This file contains the results of a bib search, hand-pruned to eliminate false hits, leaving 287 legitimate hits. The database only goes back to 1986, but my original scale space article appeared in '83, so we're missing a few of years' worth. In any event, the original citation is Andrew Witkin. Scale space filtering. In Proc. International Joint Conference on Artificial Intelligence, Karlsruhe, 1983. ------------------------------------------------------------ Author Fatemi-Ghomi, N.; Palmer, P.L.; Petrou, M.; Dept. of Electron. & Electr. Eng., Surrey Univ., Guildford, UK Title Multiresolution texture segmentation Source IEE Colloquium on 'Texture Classification: Theory and Applications (Digest No.1994/178); Part: London, UK; Part: 7 Oct. 1994; Sponsored by: IEE; IEE Colloquium on 'Texture Classification: Theory and Applications (Digest No.1994/178); London, UK; IEE; 38; 1994; pp. 2/1-3 Abstract One of the major developments in texture segmentation in recent years has been the use of multiresolution descriptions of these images. One may consider this multiresolution description to provide information about the image contained in ever smaller regions of the frequency domain, and thus provides a powerful tool for the discrimination of similar textures. It has recently been established that use of scale-space-filtering by ever wider filters, is equivalent to a decomposition of the image in terms of wavelets. The big advantage of the wavelet description of an image is that it is invertible. In this paper we use successive transformations of textured images in terms of wavelets to produce lower resolution descriptions of the image. This procedure can be thought of as dynamically focussing on the important region of the frequency domain for the particular texture in the image Thesaurus image segmentation; image texture Other Terms texture segmentation; multiresolution descriptions; frequency domain; scale-space-filtering; successive transformations; textured images ClassCodes C1250; C5260B Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4824740 AbstractNos. C9412-1250-334 References 3 Country Pub. UK date 1232 ------------------------------------------------------------ Author Ji Yu Shi; Hung Tat Tsui; Dept. of Electron. Eng., Chinese Univ. of Hong Kong, Shatin, Hong Kong Title Scale space filtering by Fejer kernel Source Proceedings of ICSIPNN '94. International Conference on Speech, Image Processing and Neural Networks; Part: vol.2; Part: Hong Kong; Part: 13-16 April 1994; Sponsored by: IEEE Signal Process. Soc.; HKIE Electron. Div.; IEE Hong Kong Centre; ISSIPNN '94. 1994 International Symposium on Speech, Image Processing and Neural Networks Proceedings (Cat. No.94TH0638-7); New York, NY, USA; IEEE; 2 vol. (xxxvii+xi+797); 1994; pp. 638-41 vol.2 Abstract A multiscale description of the raw signal data must be first obtained before more abstract representations and descriptions are generated. Many unsuccessful attempts have been made to let this initial description be as compact and complete as possible. We propose a new filtering kernel called Fejer kernel to try to tackle this problem. The behaviors of signals' features in different scales are investigated in Fejer scale space-(x, Delta ) plane. Simulation results are shown to demonstrate this new multiscale description for a signal. It compares favorably with Gaussian scale space filtering in that it brings out the main features of the signal under large scales. Gaussian filtering just makes every feature smoother. Further investigation to refine this new multiscale signal description is under development Thesaurus filtering and prediction theory; signal processing Other Terms Fejer kernel; scale space filtering; signal data; multiscale description; abstract representations; filtering kernel; Fejer scale space; simulation results; Gaussian scale space filtering; multiscale signal description; signal representation ClassCodes B6140; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4824557 AbstractNos. B9412-6140-269; C9412-1260-245 ISBN or SBN 0 7803 1865 X References 12 U.S. Copyright Clearance Center Code 0 7803 1865 X/94/$3.00 Country Pub. USA date 1226 ------------------------------------------------------------ Author Wong, Y.-f.; Lawrence Livermore Nat. Lab., Livermore, CA, USA Title A clustering filter for scale-space filtering and image restoration Source Proceedings of IEEE Conference on Computer Vision and Pattern Recognition; Part: New York, NY, USA; Part: 15-17 June 1993; Sponsored by: IEEE Comput. Soc. Tech. Committee on Pattern Anal. & Mach. Intelligence; Proceedings. 1993 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.93CH3309-2); Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xviii+804; 1993; pp. 668-9 Abstract A nonlinear clustering filter is derived using the maximum entropy principle. This filter is governed by a single-scale parameter and uses local characteristics in the data to determine the scale parameter in the output space. It provides a mechanism for removing impulsive noise, preserving edges, and improving smoothing of nonimpulsive noise. It also presents a scheme for nonlinear scale-space filtering. Comparisons with Gaussian scale- space filtering are made using real images. It is demonstrated that the clustering filter gives much better results Thesaurus entropy; filtering and prediction theory; image recognition; image reconstruction; parameter estimation Other Terms impulsive noise removal; edge preservation; nonimpulsive noise smoothing; image restoration; nonlinear clustering filter; maximum entropy principle; single-scale parameter; scale parameter; output space; nonlinear scale-space filtering ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4823739 AbstractNos. B9412-6140C-424; C9412-1250-282 ISBN or SBN 0 8186 3880 X References 5 U.S. Copyright Clearance Center Code 1063-6919/93/$03.00 Country Pub. USA date 1215 ------------------------------------------------------------ Author Mokhtarian, F.; NTT Basic Res. Lab., Tokyo, Japan Title Multi-scale, torsion-based shape representations for space curves Source Proceedings of IEEE Conference on Computer Vision and Pattern Recognition; Part: New York, NY, USA; Part: 15-17 June 1993; Sponsored by: IEEE Comput. Soc. Tech. Committee on Pattern Anal. & Mach. Intelligence; Proceedings. 1993 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.93CH3309-2); Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xviii+804; 1993; pp. 660-1 Abstract Space curves are efficient, highly descriptive features of 3-D objects. A new multiscale shape representation (the resampled torsion scale space or TSS image) for space curves suitable for recognition of a noisy curve at any scale or orientation is introduced. A two-phase matching algorithm consisting of TSS matching (using extrema of TSS zero-crossing contours) followed by transformation parameter optimization demonstrates the usefulness of the representation in recognition tasks Thesaurus image recognition; image sequences Other Terms noisy curve recognition; zero-crossing contour extrema; torsion- based shape representations; space curves; 3-D objects; multiscale shape representation; resampled torsion scale space; TSS image; two-phase matching algorithm; transformation parameter optimization ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4823735 AbstractNos. B9412-6140C-421; C9412-1250-278 ISBN or SBN 0 8186 3880 X References 2 U.S. Copyright Clearance Center Code 1063-6919/93/$03.00 Country Pub. USA date 1215 ------------------------------------------------------------ Author Lions, P.-L.; CEREMADE, Paris IX Univ., France Title Axiomatic derivation of image processing models Source Mathematical Models & Methods in Applied Sciences; Math. Models Methods Appl. Sci. (Singapore); vol.4, no.4; Aug. 1994; pp. 467-75 Abstract We briefly review the derivation due to Alvarez, Guichard, Morel (1993) and the author of mathematical models in image processing. We deduce from classical axioms in computer vision some nonlinear partial differential equations of evolution type that correspond to general multi-scale analysis (scale-space). We also obtain specific nonlinear models that satisfy additional invariances which are relevant for the analysis of images Thesaurus computer vision; nonlinear differential equations; partial differential equations Other Terms image processing models; computer vision; nonlinear partial differential equations; multi-scale analysis; scale-space ClassCodes C1250; C5260B; C4170 Article Type Theoretical / Mathematical Coden MMMSEU Language English RecordType Journal ControlNo. 4822130 AbstractNos. C9412-1250-230 ISSN 02182025 References 24 Country Pub. Singapore date 1230 ------------------------------------------------------------ Author Theimer, W.M.; Mallot, H.A.; Inst. fur Neuroinf., Ruhr-Univ., Bochum, Germany Title Phase-based binocular vergence control and depth reconstruction using active vision Source CVGIP: Image Understanding; CVGIP, Image Underst. (USA); vol.60, no.3; Nov. 1994; pp. 343-58 Abstract We present a technique for guiding vergence movements for an active stereo camera system and for calculating dense disparity maps. Both processes are described in the same theoretical framework based on phase differences in complex Gabor filter responses, modeling receptive field properties in the visual cortex. While the camera movements are computed with input images of coarse spatial resolution, the disparity map calculation uses a finer resolution in the scale space. The correspondence problem is solved implicitly by restricting the disparity range around zero disparity (Panum's area in the human visual system). The vergence process is interpreted as a mechanism to minimize global disparity, thereby setting a 3D region of interest for subsequent disparity detection. The disparity map represents smaller local disparities as an important cue for depth perception. Experimental data for the integrated performance of vergence in natural scenes followed by disparity map calculations are presented Thesaurus computer vision; stereo image processing; visual perception Other Terms phase-based binocular vergence control; depth reconstruction; active vision; guiding vergence movements; active stereo camera system; dense disparity maps; phase differences; complex Gabor filter responses; receptive field properties; visual cortex; camera movements; input images; spatial resolution; correspondence problem; human visual system; Panum's area; zero disparity; 3D region; depth perception; integrated performance; natural scenes ClassCodes B6140C; C5260B Article Type Theoretical / Mathematical Coden CIUNEJ Language English RecordType Journal ControlNo. 4822092 AbstractNos. B9412-6140C-361; C9412-5260B-232 ISSN 10499660 References 26 U.S. Copyright Clearance Center Code 1049-9660/94/$6.00 Country Pub. USA date 1233 ------------------------------------------------------------ Author Hwang, W.-L.; Mallat, S.; Dept. of Comput. Sci., New York Univ., NY, USA Title Characterization of self-similar multifractals with wavelet maxima Source Applied and Computational Harmonic Analysis; Appl. Comput. Harmon. Anal. (USA); vol.1, no.4; Sept. 1994; pp. 316-28 Abstract Self-similar multifractals have a wavelet transform whose maxima define self-similar curves in the scale-space plane. We introduce an algorithm that recovers the affine self-similarity parameters through a voting procedure in the corresponding parameter space. The voting approach is robust with respect to renormalization noise and can recover the value of parameters having random fluctuations. We describe numerical applications to Cantor measures, dyadic multifractals, and the study of diffusion limited aggregates Thesaurus fractals; wavelet transforms Other Terms self-similar multifractals; wavelet maxima; wavelet transform; self-similar curves; scale-space plane; affine self-similarity parameters; voting procedure; parameter space; renormalization noise; random fluctuations; Cantor measures; dyadic multifractals; diffusion limited aggregates ClassCodes B0230; B0290Z; C1130; C4190 Article Type Theoretical / Mathematical Coden ACOHE9 Language English RecordType Journal ControlNo. 4812736 AbstractNos. B9412-0230-002; C9412-1130-001 ISSN 10635203 References 13 U.S. Copyright Clearance Center Code 1063-5203/94/$6.00 Country Pub. USA date 1231 ------------------------------------------------------------ Author Jang, B.K.; Health Sci. Res. Lab., Eastman Kodak Co., Rochester, NY, USA Title Gaussian and morphological scale space for shape analysis of medical images Source 1993 IEEE Conference Record Nuclear Science Symposium and Medical Imaging Conference; Part: vol.2; Part: San Francisco, CA, USA; P art: 31 Oct.-6 Nov. 1993; 1993 IEEE Conference Record. Nuclear Science Symposium and Medical Imaging Conference (Cat. No.93CH3374-6); New York, NY, USA; IEEE; 3 vol. 1930; 1993; pp. 1327-31 vol.2 Editor Klaisner, L. Abstract The construction of scale space requires the smoothing of the given image to generate a set of corresponding images at other coarser scales, and the extraction of features at these scales. Various methods of smoothing, combined with various feature extractors, will result in drastically different scale space representations. The particular application and desired criteria determine the choice of smoothing and feature extractor. Here the author first puts forward a framework for the scale space representation. Then he focuses on the Gaussian and morphological scale space for planar shape analysis of medical images. The discussion of the various scale space methods is organized into three categories - boundary approach, region approach, and hybrid approach. Properties, limitations, performance, and applications of these scale space methods are discussed Thesaurus feature extraction; image segmentation; medical image processing Other Terms morphological scale space; medical images shape analysis; Gaussian space; planar shape analysis; boundary approach; region approach; hybrid approach; feature extractor; smoothing ClassCodes A8710; A8760; A8770E Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4811940 AbstractNos. A9424-8710-004 ISBN or SBN 0 7803 1487 5 References 13 U.S. Copyright Clearance Center Code 0 7803 1487 5/94/$04.00 Country Pub. USA date 1219 ------------------------------------------------------------ Author Topkar, V.; Sood, A.K.; Kjell, B.; Dept. of Comput. Sci., George Mason Univ., Fairfax, VA, USA Title Performance analysis of 1-D scale-space algorithms for pulse detection in noisy image scans Source CVGIP: Image Understanding; CVGIP, Image Underst. (USA); vol.60, no.2; Sept. 1994; pp. 191-209 Abstract Scale-space representation is a topic of active research in computer vision. The focus of the research so far has been on coarse-to-line focusing methods, image reconstruction, and computational aspects. However, not much work has been done on the signal detection problem, i.e., detecting the presence or absence of signal models from noisy image scans using the scale- space. In this paper we propose four 1-D signal detection algorithms for separating pulse signals in an image scan from the background in the scale-space domain. These algorithms do not need any thresholding to detect the zero-crossings (zc's) at any of the scales Thesaurus computer vision; image reconstruction; noise; signal detection Other Terms 1-D scale-space algorithms; pulse detection; noisy image scans; computer vision; image reconstruction; signal detection; zero- crossings; image scan ClassCodes B6140C; C5260B; C1250 Article Type Practical Coden CIUNEJ Language English RecordType Journal ControlNo. 4803851 AbstractNos. B9412-6140C-107; C9412-5260B-064 ISSN 10499660 References 17 U.S. Copyright Clearance Center Code 1049-9660/94/$6.00 Country Pub. USA date 1231 ------------------------------------------------------------ Author Kelch, J.; Wein, B.; Rogowski-Inst. for Electr. Eng., Tech. Hochschule Aachen, Germany Title Segmentation of the tongue surface in ultrasonic images using modified scale space filtering Source Proceedings of IEEE Ultrasonics Symposium; Part: vol.2; Part: Baltimore, MD, USA; Part: 31 Oct.-3 Nov. 1993; Sponsored by: Ultrasonics Ferroelectric & Frequency Control Soc; IEEE 1993 Ultrasonics Symposium Proceedings (Cat. No.93CH3301-9); New York, NY, USA; IEEE; 2 vol. 1270; 1993; pp. 947-50 vol.2 Editor Levy, M.; McAvoy, B.R. Abstract For the detection of motility disturbances during swallowing and articulation B-mode sonography is used. In general these images of a median sagittal plane through the tongue and the floor of the mouth are manually contoured by the physician, because it is quite difficult to get subtle information concerning tongue movement and motility from the real-time video images. The digitised images built a spatiotemporal space with a depth of about 70 frames. The authors' application demonstrates a way to extract the tongue contour automatically. They use a modified scale space filter for edge detection. In this way contour segments of the tongue dorsum and other object parts are extracted. A model of the tongue supports identification of the tongue segments and interpolation of the plane in the spatiotemporal space. A virtual three-dimensional view of these contours of the spatiotemporal space assists the physician to a diagnosis of the movement disturbance Thesaurus acoustic signal processing; biomedical ultrasonics; edge detection; filtering and prediction theory; image segmentation; medical image processing; patient diagnosis Other Terms tongue surface; ultrasonic images; modified scale space filtering; motility disturbances; swallowing; articulation; B- mode sonography; median sagittal plane; mouth; tongue movement; spatiotemporal space; digitised images; tongue contour; edge detection; virtual three-dimensional view; diagnosis ClassCodes A8760B; A8770E; A4360; B7510B; B7820; B6140C; C7330; C1250; C1260; C5260B Article Type Practical Language English RecordType Conference ControlNo. 4800558 AbstractNos. A9423-8760B-012; B9412-7510B-011; C9412-7330-020 ISBN or SBN 0 7803 1278 3 References 7 U.S. Copyright Clearance Center Code 1051-0117/93/$4.00 Country Pub. USA date 1219 ------------------------------------------------------------ ------------------------------------------------------------ Author Lefevre, M.; Preteux, F.; Lavayssiere, B.; Dpt. Image, ENST, Paris, France Title Radiographic inspection, defects segmentation and multiscale analysis Source Image Algebra and Morphological Image Processing V; Part: San Diego, CA, USA; Part: 25-26 July 1994; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2300; 1994; pp. 122-32 Abstract In radiographic inspection used by Electricite De France (EDF) for pipe control in nuclear power plants, pipe radiographs are digitized according to a well-defined protocol. The aim of EDF consists of developing a non-destructive testing system for identifying defects. Because of the poor quality of the digitized images, the segmentation problem is very difficult to solve. In this paper, we propose an approach based on a multiscale analysis to simultaneously achieve defect-enhancement and noise-filtering and perform accurate defect segmentation using mathematical morphology. The principle consists of defining a one-parameter family of operators which are successively applied to the digitized images after drift-removal. Equivalently, the multiscale analysis can be expressed as the solution of an evolution equation, whose analytical form is conditioned by the specific properties of the operators (e.g. mean curvature motion, affine scale-space). This results in a family of images with increasing smoothness and consistent edge preservation. Within this framework, we propose a global critical scale estimation via a multiscale analysis of second order local moments of the image. This value minimizes noise ratio without altering edges. Following the filtering step, the final segmentation is easily performed via a morphological approach based on the connection cost concept. The developed method has been tested on a set of radiographs including various defect patterns, leading to impressive segmentation results. Robustness with respect to noise ratio and optimal scale are discussed Thesaurus edge detection; flaw detection; image segmentation; mathematical morphology; noise; parameter estimation; radiography Other Terms radiographic inspection; defects segmentation; multiscale analysis; Electricite De France; pipe control; nuclear power plants; EDF; non-destructive testing system; digitized images; edge fuzziness; defect-enhancement; noise-filtering; mathematical morphology; drift-removal; evolution equation; analytical form; mean curvature motion; affine scale-space; consistent edge preservation; global critical scale estimation; second order local moments; final segmentation ClassCodes B6140C; B0590; C1250; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4791799 AbstractNos. B9411-6140C-528; C9411-1250-357 ISSN 0277786X References 25 U.S. Copyright Clearance Center Code 0 8194 1624 X/94/$6.00 Country Pub. USA date 1229 ------------------------------------------------------------ Author Peters, R.A., II; Nichols, J.A.; Dept. of Electr. Eng., Vanderbilt Univ., Nashville, TN, USA Title Properties of image sequences generated through opening residuals Source Image Algebra and Morphological Image Processing V; Part: San Diego, CA, USA; Part: 25-26 July 1994; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2300; 1994; pp. 57-68 Abstract A type of Euclidean granulometry is the sequence of images created through the opening of a single image with a family of mutually open structuring elements. The sequence is a morphological analogue of linear scale space. In such a granulometry, the minimum size of features increases in successive images, whereas in scale space the maximum spatial frequency decreases, and the images become increasingly lowpass. The opening spectrum is the image sequence created by computing the differences between successive images in a Euclidean granulometry. Whereas the difference sequence from scale space has a Laplacian bandpass characteristic, the opening spectrum does not have analogous size band-limited properties. Moreover, the images in an opening spectrum are usually not open. One can create a size band-limited sequence analogous to a Laplacian sequence through recursive opening of the tophat - the residual difference between the image and its opening. The recursion is to open the tophat with a smaller structuring element then to subtract that opening from the tophat. We call the result the tophat spectrum. This paper states and proves a number of properties of the tophat and the tophat spectrum. These include: the tophat is antiextensive and idempotent (but not increasing); each image in the tophat spectrum is size-limited and open; the structuring element family need not be mutually open to generate a tophat spectrum; if the SE family is mutually open, and the original image is binary, each image in the tophat spectrum includes the open part of the corresponding image from the opening spectrum; and, the tophat spectrum is identical to the opening spectrum created with a family of flat, one-dimensional structuring elements Thesaurus image sequences; mathematical morphology; transforms Other Terms image sequences; opening residuals; Euclidean granulometry; single image; open structuring elements; morphological analogue; linear scale space; feature size; scale space; maximum spatial frequency; opening spectrum; successive images; difference sequence; tophat; size band-limited sequence; recursive opening; antiextensive; residual difference; tophat spectrum; idempotent; size-limited image; flat one-dimensional structuring elements ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4791793 AbstractNos. B9411-6140C-522; C9411-1250-352 ISSN 0277786X References 17 U.S. Copyright Clearance Center Code 0 8194 1624 X/94/$6.00 Country Pub. USA date 1229 ------------------------------------------------------------ Author Bijaoui, A.; Obs. de la Cote d'Azur, Nice, France Title Partial reconstruction and astronomical image inventory by the wavelet transform Source Proceedings of 7th International Conference on Image Analysis and Processing; Part: Monopoli, Italy; Part: 20-22 Sept. 1993; Proceedings of the 7th International Conference on Image Analysis and Processing. Progress in Image Analysis and Processing III; Singapore; World Scientific; xxiv+733; 1994; pp. 411-16 Editor Impedovo, S. Abstract We need a vision model in order to process large astronomical images. Classical models, used on many sets of images, failed to bring a complete analysis because they are based on a single scale for the adapted smoothing and for the background mapping. That lead us to introduce a new vision model based on the wavelet transform, which splits the image in scale-space, allowing us to detect objects of different sizes. The discrete wavelet transform is obtained by the so-called "a trous" algorithm. We choose the wavelet function resulting from the difference between two successive B-spline approximations. We define an object as a subtree resulting from the image segmentation in wavelet space. The main steps of the vision procedure are: (i) the wavelet transform of the image, (ii) scale/scale thresholding of the wavelet images, (iii) field labelling of the wavelet images, and (iv) object extraction. We have built a partial reconstruction algorithm which provides an image for each object. It is easy to compute any desired parameter from any object. Our experiments showed that the detection quality is very good, and this method showed low dispersion for photometric measurements Thesaurus astronomical photometry; astronomy computing; computer vision; image reconstruction; image segmentation; wavelet transforms Other Terms partial image reconstruction; astronomical image inventory; discrete wavelet transform; vision model; adapted smoothing; background mapping; scale-space; astronomical object detection; a trous algorithm; successive B-spline approximations; subtree; image segmentation; scale/scale thresholding; field labelling; object extraction; detection quality; dispersion; photometric measurements ClassCodes A9575M; A9575P; A9575D; A4230V; C7350; C5260B; C1130 Article Type Practical; Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4784587 AbstractNos. A9422-9575-006; C9411-7350-011 ISBN or SBN 981 02 1552 5 References 14 Country Pub. Singapore date 1218 ------------------------------------------------------------ Author Sapiro, G.; Bruckstein, A.M.; Dept. of Electr. Eng., Technion-Israel Inst. of Technol., Haifa, Israel Title A B-spline based affine invariant multiscale shape representation Source Proceedings of 7th International Conference on Image Analysis and Processing; Part: Monopoli, Italy; Part: 20-22 Sept. 1993; Proceedings of the 7th International Conference on Image Analysis and Processing. Progress in Image Analysis and Processing III; Singapore; World Scientific; xxiv+733; 1994; pp. 137-44 Editor Impedovo, S. Abstract Multiscale representations constitute an important topic in different fields as computer vision, CAD, and image processing. In this work, a multiscale representation of planar shapes is described. The approach is based on computing B-splines of increasing orders, and therefore is automatically affine invariant. The resulting representation satisfies all the basic scale-space properties at least in a qualitative form, and is simple to implement Thesaurus CAD; computer vision; splines (mathematics) Other Terms B-spline based affine invariant multiscale shape representation; computer vision; CAD; image processing; planar shapes; scale- space properties ClassCodes B6140C; B0290F; C5260B; C4130 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 4784549 AbstractNos. B9411-6140C-290; C9411-5260B-232 ISBN or SBN 981 02 1552 5 References 23 Country Pub. Singapore date 1218 ------------------------------------------------------------ Author Theimer, W.M.; Mallot, H.A.; Inst. fur Neuroinf., Ruhr-Univ., Bochum, Germany Title Vergence guided depth reconstruction using a phase method Source Proceedings of Sixth International Conference on Neural Networks and their Industrial and Cognitive Applications; Part: Nimes, France; Part: 25-29 Oct. 1993; Sponsored by: AFIA; ARC; IFIP; JSAI & SEE; EERIE; et al; Sixth International Conference. Neural Networks and their Industrial and Cognitive Applications. NEURO-NIMES 93 Conference Proceedings and Exhibition Catalog; Nanterre, France; EC2; 500; 1993; pp. 299-308 Abstract We present a technique to guide vergence movements for an active stereo camera system and to calculate dense disparity maps. Both processes are described in the same theoretical framework based on phase differences in complex Gabor filter responses, modelling receptive field properties in the visual cortex. While the camera movements are computed with input images of coarse spatial resolution, disparity calculation uses a finer resolution in the scale space. The correspondence problem is solved implicitly by restricting the disparity range around zero disparity to the filter kernel sizes (Panum's area in the human visual system). The vergence process is interpreted as a mechanism to minimize global disparity, thereby setting a 3D region of interest for subsequent disparity detection. The disparity map represents smaller local disparities as an important cue for depth perception. Experimental data for vergence in natural scenes and resulting disparity maps are presented Thesaurus image reconstruction; stereo image processing Other Terms depth reconstruction; stereo camera system; dense disparity maps ; Gabor filter responses; visual cortex; disparity calculation; correspondence problem; disparity range; global disparity; natural scenes ClassCodes C5260B; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4778274 AbstractNos. C9411-5260B-195 ISBN or SBN 2 906899 95 X References 12 Country Pub. France date 1219 ------------------------------------------------------------ Author Falzon, F.; Giraudon, G.; INRIA, Sophia-Antipolis, France Title Singularity analysis and derivative scale-space Source Proceedings of IEEE Conference on Computer Vision and Pattern Recognition; Part: Seattle, WA, USA; Part: 21-23 June 1994; Sponsored by: IEEE Comput. Soc. Tech. Committee on Pattern Anal. & Machine Intelligence; Proceedings 1994 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.94CH3405-8); Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xvi+1009; 1994; pp. 245-50 Abstract The scale-space representations and wavelet transform theory have provided us with good singularity detection frameworks. Thanks to such methods, image understanding processes have become more and more powerful. Some computer vision tasks also require a knowledge on the nature of detected singularities. We propose, in this paper, to take into account these requirements, and we present a singularity detection method based on fractional calculus arguments Thesaurus computer vision; wavelet transforms Other Terms scale-space representations; wavelet transform; singularity detection; image understanding; computer vision; fractional calculus ClassCodes C1250; C5260B; C1130 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4777938 AbstractNos. C9411-1250-133 ISBN or SBN 0 8186 5825 8 References 26 U.S. Copyright Clearance Center Code 1063-6919/94/$3.00 Country Pub. USA date 1228 ------------------------------------------------------------ Author Matsopoulos, G.; Marshall, S.; Strathclyde Univ., Glasgow, UK Title Feature migration in morphological scale space Source Proceedings of ICASSP '93; Part: Minneapolis, MN, USA; Part: 27- 30 April 1993; Sponsored by: IEEE Signal Process. Soc; ICASSP-93. 1993 IEEE International Conference on Acoustics, Speech, and Signal Processing (Cat. No.92CH3252-4); New York, NY, USA; IEEE; 5 vol. (652+735+606+559+681); 1993; pp. 599-602 vol.3 Abstract A novel morphological operator for scale space evaluation is introduced. The operator (X/sub B/)/sup B/ is less prone to object shrinkage and signal spreading than the Gaussian approach, and it treats similarly the intrusions and protrusions of the original image. Zero crossing migration with respect to this operation for various types of signal is derived. In order to track events efficiently from scale to scale, it is necessary to understand how the zero crossings of a signal migrate with the scale factor. It is shown that equations can be produced to describe this movement for specific special cases, such as ramp and edges. The new theoretical results can be used to make feature tracking with nonlinear techniques more efficient Thesaurus feature extraction; image processing; mathematical morphology; tracking Other Terms morphological operator; scale space evaluation; object shrinkage ; signal spreading; zero crossings; feature tracking; nonlinear techniques ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4777723 AbstractNos. B9411-6140C-204; C9411-1250-123 ISBN or SBN 0 7803 0946 4 References 5 U.S. Copyright Clearance Center Code 0 7803 0946 4/93/$3.00 Country Pub. USA date 1213 ------------------------------------------------------------ Author Ogniewicz, R.L.; Commun. Technol. Lab., Swiss Federal Inst. of Technol., Zurich, Switzerland Title Skeleton-space: a multiscale shape description combining region and boundary information Source Proceedings of IEEE Conference on Computer Vision and Pattern Recognition; Part: Seattle, WA, USA; Part: 21-23 June 1994; Sponsored by: IEEE Comput. Soc. Tech. Committee on Pattern Anal. & Machine Intelligence; Proceedings 1994 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.94CH3405-8); Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xvi+1009; 1994; pp. 746-51 Abstract A multiscale extension to the medial axis transform(MAT) or skeleton can be obtained by combining information derived from a scale-space hierarchy of boundary representations with region information provided by the MAT. The skeleton-space is constructed by attributing each skeleton component with a hierarchically ordered sequence of residual values, each expressing the saliency of the component at a distinct resolution level. Since our method amounts to a rather symbolic than iconic computation of a multiscale MAT, it does not introduce the correspondence problem between distinct levels of detail, in contrast to other commonly proposed techniques. Our multiscale MAT is capable of describing complex shapes characterized by significantly jagged boundaries. Furthermore, tracking the evolution of prominent loci of the MAT such as nodes across scales permits to assess the most significant skeleton constituents and to automatically determine pruning parameters. A salient subset of the MAT (first order skeleton) can be extracted without the need of manual threshold adjustment Thesaurus image processing; mathematical morphology Other Terms skeleton-space; multiscale shape description; boundary information; region information; medial axis transform; scale- space hierarchy; boundary representations; hierarchically ordered sequence; pruning parameters; image analysis; mathematical morphology ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 4771239 AbstractNos. B9411-6140C-098; C9411-1250-080 ISBN or SBN 0 8186 5825 8 References 16 U.S. Copyright Clearance Center Code 1063-6919/94/$3.00 Country Pub. USA date 1228 ------------------------------------------------------------ Author Goshtasby, A.; O'Neill, W.D.; Dept. of Electr. Eng. & Comput. Sci., Chicago Univ., IL, USA Title Curve fitting by a sum of Gaussians Source CVGIP: Graphical Models and Image Processing; CVGIP, Graph. Models Image Process. (USA); vol.56, no.4; July 1994; pp. 281-8 Abstract Gaussians are useful in multiscale representation of video data. An algorithm is presented which approximates a sequence of uniformly spaced single-valued data by a sum of Gaussians with a prescribed accuracy. The scale-space image of the data is used to estimate the number of Gaussians and their initial parameters. The Marquardt algorithm (1963) is then used to optimize the parameters Thesaurus curve fitting; image processing; optimisation; parameter estimation Other Terms curve fitting; Gaussians; multiscale representation; video data ; Marquardt algorithm; initial parameters; scale-space image ClassCodes C5260B; C1250; C1180; C6130B Article Type Theoretical / Mathematical Coden CGMPE5 Language English RecordType Journal ControlNo. 4769682 AbstractNos. C9411-5260B-027 ISSN 10499652 References 8 U.S. Copyright Clearance Center Code 1049-9652/94/$6.00 Country Pub. USA date 1229 ------------------------------------------------------------ Author Tremblay, M.; Savard, M.; Poussart, D.; Dept. of Electr. Eng., Laval Univ., Que., Canada Title Medium level scene representation using a VLSI smart hexagonal sensor with multiresolution edge extraction capability and scale space integration processing Source Proceedings of IEEE Conference on Computer Vision and Pattern Recognition; Part: Seattle, WA, USA; Part: 21-23 June 1994; Sponsored by: IEEE Comput. Soc. Tech. Committee on Pattern Anal. & Machine Intelligence; Proceedings 1994 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.94CH3405-8); Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xvi+1009; 1994; pp. 632-7 Abstract This paper presents a sensing approach where photo-transduction, multi-resolution feature extraction, scale-space integration and edge tracking are performed on a mixed (digital-analog) VLSI architecture in order to generate medium-level scene description. The proposed system is mainly targeted for robot vision applications where feature description is preferred to a set of raw or raster 2D images and edge maps. The Multiport Access photo- Receptor (MAR) is a CMOS sensor and represents the main sensory part of this integrated image acquisition system. VLSI also provides means to integrate analog computing, digital controller and DSP co-processor modules which define a powerful sensory chip set for focal plane image processing. A current version of the MAR sensor which implements 256*256 pixels includes 16 analog spatial filters which simultaneously compute multiresolution edge maps. This unique 2D hexagonal smart sensor approach which performs up to 8.5*10/sup 9/ arithmetic Op/sec during the acquisition/filtering phase and 25*10/sup 9/ Logical Op/sec for scale-space integration allows high resolution image capability. It represents a significant improvement for passive sensory units in a compact assembly for computer vision applications Thesaurus CMOS integrated circuits; computer vision; edge detection; feature extraction; image sensors; spatial filters; VLSI Other Terms medium level scene representation; VLSI smart hexagonal sensor; multiresolution edge extraction capability; scale space integration processing; feature extraction; edge tracking; mixed digital-analog VLSI architecture; robot vision; Multiport Access photo-Receptor; CMOS sensor; image acquisition system; digital controller; DSP co-processor modules; focal plane image processing; analog spatial filters; multiresolution edge maps ClassCodes A4280Q; B7230G; B6140C; B2570D; C1250; C5260B; C3240K Article Type Applications; Practical Language English RecordType Conference ControlNo. 4764238 AbstractNos. A9420-4280Q-004; B9410-7230G-033; C9410-1250-212 ISBN or SBN 0 8186 5825 8 References 12 U.S. Copyright Clearance Center Code 1063-6919/94/$3.00 Country Pub. USA date 1228 ------------------------------------------------------------ Author Jang, B.K.; Health Sci. Res. Lab., Eastman Kodak Co., Rochester, NY, USA Title Morphological scale space for shape smoothing Source Proceedings of the Fourth International Conference on Signal Processing Applications and Technology. ICSPAT '93; Part: vol.1; Part: Santa Clara, CA, USA; Part: 28 Sept.-1 Oct. 1993; Proceedings of the Fourth International Conference on Signal Processing Applications and Technology; Newton, MA, USA; DSP Associates; 2 vol. 1675; 1993; pp. 746-55 vol.1 Abstract In this paper, we describe a multiscale image representation based on morphological operations. A set of boundaries of an object is first smoothed by a number of opening and closing operations each using a different size structuring element, generating the multiple scale representation. Then, depending on specific applications, features of the smoothed boundary are extracted and linked together across a continuum of scales forming a map called the morphological scale space (MSS). Properties of this scale space map are investigated. A shape smoothing algorithm is presented to show how the scale space representation along with an organization scheme could be applied to solve some practical vision problems. Specifically, in line with Witkin's idea of Gaussian scale space filtering, the organization scheme is developed to determine the relationship among various boundary features. As a consequence, the proposed algorithm allows the selection of the perceptually dominant scales for shape smoothing without prior knowledge of the object boundary and its noise. The proposed algorithm uses both region and boundary information, addresses the issues of local and global information, and solves the problems of scale. Moreover, the method is computationally efficient, parameter free, and stable Thesaurus computer vision; feature extraction; mathematical morphology Other Terms morphological scale space; shape smoothing; multiscale image representation; closing operations; opening operations; structuring element; smoothed boundary; feature extraction; scale space map; practical vision problems; Gaussian scale space filtering; boundary features; perceptually dominant scales ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4759476 AbstractNos. B9410-6140C-309; C9410-1250-171 References 12 Country Pub. USA date 1218 ------------------------------------------------------------ Author Jones, A.G.; Taylor, C.J.; Dept. of Med. Biophys., Manchester Univ., UK Title Robust shape from shading Source Image and Vision Computing; Image Vis. Comput. (UK); vol.12, no.7; Sept. 1994; pp. 411-21 Abstract Existing shape from shading algorithms are not robust enough to work reliably with real images. We present a new paradigm for shape from shading based upon a scale space surface representation. Using this representation, the shape from shading problem becomes one of forming a surface from Gaussian basis functions in such a way that the shading in the original image is explained. The algorithm is extremely robust, producing convincing results even when there is considerable noise present. In its current form, the algorithm requires the reflectance function to be approximately circularly symmetric. We show results for synthetic images, both with and without added noise, and for real scanning electron microscope images Thesaurus computer vision; image recognition; noise Other Terms algorithms; real images; scale space surface representation; robust shape from shading; surface; Gaussian basis functions; noise; reflectance function; synthetic images; scanning electron microscope images ClassCodes C5260B Article Type Practical Coden IVCODK Language English RecordType Journal ControlNo. 4756038 AbstractNos. C9410-5260B-157 ISSN 02628856 References 15 U.S. Copyright Clearance Center Code 0262-8856/94/07/0411-11$10.00 Country Pub. UK date 1231 ------------------------------------------------------------ Author Kulkarni, A.; Shevgaonkar, R.; Sahasrabudhe, S.C.; Dept. of Electr. Eng., Indian Inst. of Technol., Bombay, India Title Edge detection using scale space knowledge Source Proceedings of TENCON '93. IEEE Region 10 International Conference on Computers, Communications and Automation; Part: vol.2; Part: Beijing, China; Part: 19-21 Oct. 1993; Proceedings TENCON '93. 1993 IEEE Region 10 Conference on 'Computer, Communication, Control and Power Engineering' (Cat. No. 93CH3286-2); New York, NY, USA; IEEE; 5 vol. (xxvi+1206+xvii+676+xv+580+xvii+619); 1993; pp. 986-90 vol.2 Editor Yuan Baozong Abstract Edge detection has acquired enormous importance in computer vision research: gaussian filter has been widely used in the past to accomplish this task. Although the gaussian filter has nice scaling behaviour and is computationally elegant, it suffers from the disadvantage of delocalisation of edges when operated at higher scales. Multi-scale processing of an image is thus required. In this paper, the process of edge detection has been looked upon as a reasoning problem. In the past, knowledge handling and reasoning have been attributed to high level vision routines, but from the results presented here, it can be argued that reasoning does play an important role in the process of edge detection. The knowledge base required for this is formed out of the theory of scale space. Particular emphasis is laid on the behaviour of delocalised, missing and false or spurious edges in the scale space. The scale space is formed by operating the LoG filter of different sizes on the input image. The dissimilarity among the zero crossings is measured across the scales. It is useful in the removal of false edges. A compatibility measure relates the zero-crossing contour with the current scale parameter, sigma . The rules are coded in NEXPERT OBJECT 2.0. The results are compared with Canny's multiple edge detection algorithm (J.F. Canny, 1986) Thesaurus computer vision; edge detection; inference mechanisms; knowledge based systems Other Terms edge detection; scale space knowledge; computer vision research; gaussian filter; scaling behaviour; multi-scale processing; image processing; reasoning problem; knowledge handling; reasoning; knowledge based systems; spurious edges; LoG filter; zero-crossing contour; NEXPERT OBJECT 2.0; multiple edge detection algorithm ClassCodes B6140C; C5260B; C6170 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 4750929 AbstractNos. B9410-6140C-204; C9410-5260B-121 ISBN or SBN 0 7803 1233 3 References 16 Country Pub. USA date 1219 ------------------------------------------------------------ Author Fermuller, C.; Kropatsch, W.; Comput. Vision Lab., Maryland Univ., College Park, MD, USA Title A syntactic approach to scale-space-based corner description Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.16, no.7; July 1994; pp. 748-51 Abstract Planar curves are described by information about corners integrated over various levels of resolution. The detection of corners takes place on a digital representation. To compensate for ambiguities arising from sampling problems due to the discreteness, results about the local behavior of curvature extrema in continuous scale-space are employed Thesaurus edge detection; image processing; parallel processing Other Terms syntactic approach; scale space based corner description; planar curves; resolution; corner detection; sampling problems; curvature extrema ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4742145 AbstractNos. B9410-6140C-083; C9410-1250-056 ISSN 01628828 References 23 U.S. Copyright Clearance Center Code 0162-8828/94/$04.00 Country Pub. USA date 1229 ------------------------------------------------------------ Author Jianqing Liu; Yee-Hong Yang; Comput. Vision Lab., Saskatchewan Univ., Saskatoon, Sask., Canada Title Multiresolution color image segmentation Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.16, no.7; July 1994; pp. 689-700 Abstract Image segmentation is the process by which an original image is partitioned into some homogeneous regions. In this paper, a novel multiresolution color image segmentation (MCIS) algorithm which uses Markov random fields (MRF's) is proposed. The proposed approach is a relaxation process that converges to the MAP (maximum a posteriori) estimate of the segmentation. The quadtree structure is used to implement the multiresolution framework, and the simulated annealing technique is employed to control the splitting and merging of nodes so as to minimize an energy function and therefore, maximize the MAP estimate. The multiresolution scheme enables the use of different dissimilarity measures at different resolution levels. Consequently, the proposed algorithm is noise resistant. Since the global clustering information of the image is required in the proposed approach, the scale space filter (SSF) is employed as the first step. The multiresolution approach is used to refine the segmentation. Experimental results of both the synthesized and real images are very encouraging. In order to evaluate experimental results of both synthesized images and real images quantitatively, a new evaluation criterion is proposed and developed Thesaurus image segmentation; Markov processes; simulated annealing; tree data structures Other Terms multiresolution color image segmentation; Markov random fields; relaxation process; maximum a posteriori estimate; quadtree structure; simulated annealing; splitting; merging; dissimilarity measures; noise resistant; global clustering information; scale space filter; real images; synthesized images; evaluation criterion ClassCodes B6140C; B0260; B0240Z; C1250; C5260B; C1180; C1140Z Article Type Practical; Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4742138 AbstractNos. B9410-6140C-078; C9410-1250-051 ISSN 01628828 References 19 U.S. Copyright Clearance Center Code 0162-8828/94/$04.00 Country Pub. USA date 1229 ------------------------------------------------------------ Author Pizer, S.M.; Burbeck, C.A.; Coggins, J.M.; Fritsch, D.S.; Morse, B.S.; Medical Image Display Res. Group, North Carolina Univ., Chapel Hill, NC, USA Title Object shape before boundary shape: scale-space medial axes Source Journal of Mathematical Imaging and Vision; J. Math. Imaging Vis. (Netherlands); vol.4, no.3; July 1994; pp. 303-13 Abstract Representing object shape in two or three dimensions has typically involved the description of the object boundary. This paper proposes a means for characterizing object structure and shape that avoids the need to find an explicit boundary. Rather, it operates directly from the image-intensity distribution in the object and its background, using operators that respond to "boundariness." It produces a sort of medial-axis description that recognizes that both axis location and object width must be defined according to a tolerance proportional to the object width. This generalized axis is called the multiscale medial axis because it is defined as a curve or set of curves in scale space. It has all of the advantages of the traditional medial axis. It also has significant new advantages: it does not require a predetermination of exactly what locations are included in the object, it provides gross descriptions that are stable against image detail, and it can be used to identify subobjects and regions of boundary detail and to characterize their shape properties Thesaurus edge detection; image processing Other Terms boundary shape; scale space medial axes; object boundary; object structure; object shape; image intensity distribution; medial axis description; multiscale medial axis; image analysis; shape description ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden JIMVEC Language English RecordType Journal ControlNo. 4739197 AbstractNos. B9410-6140C-034; C9410-1250-023 ISSN 09249907 References 26 U.S. Copyright Clearance Center Code 0924-9907/94/$5.00 Country Pub. Netherlands date 1229 ------------------------------------------------------------ Author Yoon, S.H.; Alexander, W.E.; Kim, J.H.; Dept. of Electr. & Comput. Eng., North Carolina State Univ., Raleigh, NC, USA Title Scale-invariant shape representation based on the wavelet transform Source Signal Processing, Sensor Fusion, and Target Recognition III; Part : Orlando, FL, USA; Part: 4-6 April 1994; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2232; 1994; pp. 142-53 Abstract Matching of occluded objects is a difficult problem. Moreover, the problem is more difficult when scale-invariant matching is needed. A scale invariant representation is essential for this application. In this paper, we propose using the wavelet transform of a boundary to obtain a scale-invariant representation. We use the cubic B-spline as a smoothing function of the wavelet transform since the B-spline is analytically well defined and simple to implement. We implement the fast continuous wavelet transform by using a dyadic wavelet decomposition and dilated B-splines. As a result of using the wavelet transform, we obtain boundaries at various scales while using a small number of data points. The existing scale-space image approaches are not effective for occluded object matching since they use a normalized x-axis and too many data points. We propose a new scale-invariant representation similar to the scale-space image. The representation is generated by locating zero-crossings of the curvature function of boundaries at only the scales where the number of zero-crossings is changing. We scale the x-axis for each scale instead of using the same normalization for all scales. The proposed representation is scale-invariant and appropriate for scale-invariant matching with occlusion Thesaurus image recognition; invariance; splines (mathematics); wavelet transforms Other Terms scale-invariant shape representation; dyadic wavelet decomposition; occluded objects; scale-invariant matching; cubic B-spline; smoothing function; fast continuous wavelet transform; dilated B-splines; scale-space image approaches; curvature function ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4737080 AbstractNos. B9410-6140C-007; C9410-1250-005 ISSN 0277786X References 21 U.S. Copyright Clearance Center Code 0 8194 1536 7/94/$6.00 Country Pub. USA date 1226 ------------------------------------------------------------ Author ter Haar Romeny, B.M.; Florack, L.M.J.; Salden, A.H.; Viergever, M.A.; 3D Comput. Vision Res. Group, Utrecht Univ. Hospital, Netherlands Title Higher order differential structure of images Source Image and Vision Computing; Image Vis. Comput. (UK); vol.12, no.6; July-Aug. 1994; pp. 317-25 Abstract This paper is meant as a tutorial on the basic concepts for vision in the 'Koenderink' school. The concept of scale-space is a necessity, if the extraction of structure from measured physical signals (i.e. images) is at stage. The Gaussian derivative kernels for physical signals are then the natural analogues of the mathematical differential operators. This paper discusses some interesting properties of the Gaussian derivative kernels, like their orthogonality and behaviour with noisy input data. Geometrical structure to extract is expressed in terms of differential invariants, in this paper limited to invariants under orthogonal transformations. Three representations are summarized: cartesian, gauge and manifest invariant notation. Many explicit examples are given. A section is included about the computer implementation of the calculation of higher order invariant structure Thesaurus computer vision; feature extraction Other Terms higher order differential image structure; measured physical signals; Gaussian derivative kernels; mathematical differential operators; manifest invariant notation; computer implementation; higher order invariant structure ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical Coden IVCODK Language English RecordType Journal ControlNo. 4727494 AbstractNos. B9409-6140C-246; C9409-1250-155 ISSN 02628856 References 36 U.S. Copyright Clearance Center Code 0262-8856/94/06/0317-09$10.00 Country Pub. UK date 1229 ------------------------------------------------------------ Author Worring, M.; Smeulders, A.W.M.; Dept. of Math. & Comput. Sci., Amsterdam Univ., Netherlands Title Shape of an arbitrary finite point set in IR/sup 2/ Source Journal of Mathematical Imaging and Vision; J. Math. Imaging Vis. (Netherlands); vol.4, no.2; May 1994; pp. 151-70 Abstract Studies the shape of a finite point set in IR/sup 2/, where the points are not bound to a regular grid like Z/sup 2/. The shape of a connected point set in IR/sup 2/ is captured by its boundary. For a finite point set the boundary is a directed graph that connects points identified as boundary points. The authors argue that to serve as a proper boundary definition the directed graph should regulate scale, be minimal, have an increasing interior and be consistent with the boundary definition of connected objects. The authors propose to use the directed variant of the alpha -shape as defined by Edelsbrunner et al. (1983), which the authors call the alpha -graph. The alpha -graph is based on a generalization of the convex hull. The computational aspects of the alpha -graph have been extensively studied, but little attention has been paid to the potential use of the alpha -graph as a shape descriptor or a boundary definition. In this paper the authors prove that the alpha -graph satisfies the aforementioned criteria. The authors also prove a relation between the alpha - graph and the opening scale space from mathematical morphology. In fact, the alpha -hull provides a generalization of this scale space Thesaurus computational geometry; directed graphs; image processing; mathematical morphology Other Terms arbitrary finite point set; IR/sup 2/; connected point set; directed graph; boundary definition; alpha -shape; alpha - graph; convex hull; shape descriptor; mathematical morphology ClassCodes B6140C; B0250; C1250; C1160; C5260B Article Type Practical; Theoretical / Mathematical Coden JIMVEC Language English RecordType Journal ControlNo. 4724559 AbstractNos. B9409-6140C-218; C9409-1250-133 ISSN 09249907 References 16 U.S. Copyright Clearance Center Code 0924-9907/94/$5.00 Country Pub. Netherlands date 1227 ------------------------------------------------------------ Author El-Fallah, A.I.; Ford, G.E.; CIPIC, California Univ., Davis, CA, USA Title Nonlinear adaptive image filtering based on inhomogeneous diffusion and differential geometry Source Image and Video Processing II; Part: San Jose, CA, USA; Part: 7- 9 Feb. 1994; Sponsored by: SPIE; IS&T; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2182; 1994; pp. 49-63 Abstract The inadequacy of the classic linear approach to edge detection and scale space filtering lies in the spatial averaging of the Laplacian. The Laplacian is the divergence of the gradient and thus is the divergence of both magnitude and direction. The divergence in magnitude characterizes edges and this divergence must not be averaged if the image structure is to be preserved. We introduce a new nonlinear filtering theory that that only averages the divergence of direction. This averaging keeps edges and lines intact as their direction is nondivergent. Noise does not have this nondivergent consistency and its divergent direction will be averaged. Higher order structures such as corners are singular points or inflection points in the divergence of direction and also will be averaged. Corners are intersection points of edges of nondivergent direction (or smooth curves of small divergence in direction) and their averaging will be limited. This approach provides a better compromise between noise removal and preservation of image structure. Experiments that verify and demonstrate the adequacy of this new theory are presented Thesaurus adaptive filters; differential geometry; edge detection; filtering and prediction theory; noise Other Terms nonlinear adaptive image filtering; inhomogeneous diffusion; differential geometry; edge detection; scale space filtering; nonlinear filtering theory; divergence of direction; nondivergent consistency; corners; singular points; inflection points; noise removal; image structure preservation; spatial averaging ClassCodes B6140C; C5260B; C1250 Article Type Theoretical / Mathematical; Experimental Coden PSISDG Language English RecordType Conference ControlNo. 4716779 AbstractNos. B9409-6140C-077; C9409-5260B-054 ISSN 0277786X References 16 U.S. Copyright Clearance Center Code 0 8194 1477 8/94/$6.00 Country Pub. USA date 1224 ------------------------------------------------------------ Author Peters, R.A., II; Nichols, J.A.; Dept. of Electr. Eng., Vanderbilt Univ., Nashville, TN, USA Title Morphological bandpass decomposition of images Source Nonlinear Image Processing V; Part: San Jose, CA, USA; Part: 7- 9 Feb. 1994; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2180; 1994; pp. 163-74 Abstract The concept of a morphological size distribution is well known. It can be envisioned as a sequence of progressively more highly smoothed images which is a nonlinear analogue of scale space. Whereas the differences between Gaussian lowpass filtered images in scale space form a sequence of approximately Laplacian bandpass filtered images, the difference image sequence from a morphological size distribution is not bandpass in any usual sense for most images. This paper presents a proof that a strictly size band limited sequence can be created along one dimension in an n dimensional image. This result is used to show how an image time sequence can be decomposed into a set of sequences each of which contains only events of a specific limited duration. It is shown that this decomposition can be used for noise reduction. This paper also presents two algorithms which create from morphological size distributions, (pseudo) size bandpass decompositions in more than one dimension. One algorithm uses Vincent grayscale reconstruction on the size distribution. The other reconstructs the difference image sequence Thesaurus image reconstruction; image sequences; mathematical morphology; random noise; statistical analysis Other Terms morphological bandpass decomposition; morphological size distribution; smoothed image sequences; scale space; difference image sequence; band limited sequence; image time sequence; noise reduction; bandpass decompositions; Vincent grayscale reconstruction; size distribution ClassCodes B6140C; C5260B Article Type Theoretical / Mathematical; Experimental Coden PSISDG Language English RecordType Conference ControlNo. 4716757 AbstractNos. B9409-6140C-057; C9409-5260B-037 ISSN 0277786X References 8 U.S. Copyright Clearance Center Code 0 8194 1475 1/94/$6.00 Country Pub. USA date 1224 ------------------------------------------------------------ Author Bangham, J.A.; Sch. of Inf. Syst., East Anglia Univ., Norwich, UK Title Median and morphological scale space filtering and zero-crossings Source Nonlinear Image Processing V; Part: San Jose, CA, USA; Part: 7- 9 Feb. 1994; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2180; 1994; pp. 90-8 Abstract This paper investigates nonlinear filter sequences (filters in series, known as sieves) or banks (in parallel) as a means of achieving multiscale decomposition. Recently multiscale decomposition using both erosion (dilation) and closing (opening) operations with sets of increasing scale flat structuring elements have been used to analyse edges over multiple scales and the granularity of images. These do not introduce new edges as scale increases. However, they are not at all statistically robust in the face of, for example, salt and pepper noise. This paper shows that sieves also do not introduce new edges, are very robust and perform at least as well as discreet Gaussian filters when applied to sampled data. Analytical support for these observations is provided by the morphology decomposition theorem Thesaurus edge detection; filtering and prediction theory; mathematical morphology; nonlinear systems Other Terms morphological scale space filtering; median scale space filtering ; zero-crossings; multiscale decomposition; scale space channels; nonlinear filter sequences; erosion operations; dilation operations; closing operation; opening operations; edge analysis; salt and pepper noise; sampled data; morphology decomposition theorem ClassCodes B6140C; C5260B; C1250 Article Type Experimental Coden PSISDG Language English RecordType Conference ControlNo. 4716750 AbstractNos. B9409-6140C-051; C9409-5260B-030 ISSN 0277786X References 15 U.S. Copyright Clearance Center Code 0 8194 1475 1/94/$6.00 Country Pub. USA date 1224 ------------------------------------------------------------ Author Miller, E.L.; Willsky, A.S.; Lab. for Inf. & Decision Syst., MIT, Cambridge, MA, USA Title Wavelet transforms and multiscale estimation techniques for the solution of multisensor inverse problems Source Wavelet Applications; Part: Orlando, FL, USA; Part: 5-8 April 1994; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2242; 1994; pp. 28-39 Abstract The application of multiscale and stochastic techniques to the solution of linear inverse problems is presented. This approach allows for the explicit and easy handling of a variety of difficulties commonly associated with problems of this type. Regularization is accomplished via the incorporation of prior information in the form of a multiscale stochastic model. We introduce the relative error covariance matrix (RECM) as a tool for quantitatively evaluating the manner in which data contributes to the structure of a reconstruction. In particular, the use of a scale space formulation is ideally suited to the fusion of data from several sensors with differing resolutions and spatial coverage (e.g. sparse or limited availability). Moreover, the RECM both provides us with an ideal tool for understanding and analyzing the process of multisensor fusion and allows us to define the space-varying optimal scale for reconstruction as a function of the nature (resolution, quality, and coverage) of the available data. Examples of our multiscale maximum posteriori inversion algorithm are demonstrated using a two channel deconvolution problem Thesaurus estimation theory; inverse problems; matrix algebra; sensor fusion; stochastic processes; wavelet transforms Other Terms multiscale estimation techniques; multisensor inverse problems; stochastic techniques; linear inverse problems; regularization; multiscale stochastic model; relative error covariance matrix; RECM; scale space formulation; data fusion; multisensor fusion; space-varying optimal scale for reconstruction; multiscale maximum posteriori inversion algorithm; channel deconvolution problem ClassCodes B6140; B0210; C1260; C1110 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4707528 AbstractNos. B9408-6140-144; C9408-1260-111 ISSN 0277786X References 10 U.S. Copyright Clearance Center Code 0 8194 1546 4/94/$6.00 Country Pub. USA date 1226 ------------------------------------------------------------ Author Bala, J.W.; Wechsler, H.; Dept. of Comput. Sci., George Mason Univ., Fairfax, VA, USA Title Learning to detect targets using scale-space and genetic search Source Proceedings of ICGA-93: Fifth International Conference on Genetic Algorithms; Part: Urbana-Champaign, IL, USA; Part: 17-22 July 1993; Proceedings of the Fifth International Conference on Genetic Algorithms; San Mateo, CA, USA; Morgan Kaufmann; xii+665; 1993; pp. 516-22 Editor Forrest, S. Abstract Introduces a novel methodology for target detection using scale- space and genetic algorithms (GAs). High-performance detection operators, defined as 2D masks, are derived using GAs operating at different levels of the Gaussian pyramid. The population of operators, iteratively evaluated according to a statistical performance index corresponding to target recognition ability, evolves into an optimal set of scale-space masks using the evolution principles of genetic search. Experimental results of target detection in poor quality and noisy infrared images are presented to illustrate the merit and feasibility of the proposed method Thesaurus genetic algorithms; image recognition; iterative methods; learning (artificial intelligence); performance index; search problems Other Terms learning; scale-space image representation; high performance detection operators; noisy IR images; iterative evaluation; poor quality images; genetic search; target detection; genetic algorithms; 2D masks; Gaussian pyramid; statistical performance index; target recognition ability; evolution ClassCodes C1250; C5260B; C1180; C1240 Article Type Practical; Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4706214 AbstractNos. C9408-1250-105 ISBN or SBN 1 558 60299 2 References 8 Country Pub. USA date 1216 ------------------------------------------------------------ Author De Graaf, C.N.; Koster, A.S.E.; Vincken, K.L.; Viergever, M.A.; Comput. Vision Res. Group, Utrecht Univ., Netherlands Title Validation of the interleaved pyramid for the segmentation of 3D vector images Source Pattern Recognition Letters; Pattern Recognit. Lett. (Netherlands); vol.15, no.5; May 1994; pp. 469-75 Abstract A multiresolution pyramid with double scale space sampling (compared to the Burt & Hong scheme, 1981) for the segmentation of 3D images, of which the elements are multiple valued, is described. Evaluation is carried out by quality constrained cost analysis Thesaurus image segmentation Other Terms interleaved pyramid; 3D vector image segmentation; multiresolution pyramid; double scale space sampling; multiple- valued images; quality constrained cost analysis ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PRLEDG Language English RecordType Journal ControlNo. 4697507 AbstractNos. B9408-6140C-054; C9408-1250-032 ISSN 01678655 References 10 U.S. Copyright Clearance Center Code 0167-8655/94/$07.00 Country Pub. Netherlands date 1227 ------------------------------------------------------------ Author Prasad, K.V.; Balaram, J.; California Inst. of Technol., Pasadena, CA, USA Title Automated inspection for remote telerobotic operations Source Proceedings of 1993 IEEE International Conference on Robotics and Automation; Part: Atlanta, GA, USA; Part: 2-6 May 1993; Sponsored by: IEEE; Proceedings IEEE International Conference on Robotics and Automation (Cat. No.93CH3247-4); Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 3 vol. (xviii+1051+xvi+848+xviii+1042); 1993; pp. 883-8 vol.3 Abstract Automated inspection techniques being developed for surface inspection of remote space platforms are reported. The unique problems of performing visual telerobotic inspection in space are identified. An image differencing method to detect changes to surfaces over a period of time is presented, together with a scale-space technique for flaw identification. Examples from images of laboratory mockups of space platform modules are presented to illustrate the results Thesaurus aerospace control; automatic optical inspection; mobile robots; space vehicles; telecontrol Other Terms remote telerobotic operations; surface inspection; space platforms; visual telerobotic inspection; image differencing method; scale-space technique; flaw identification ClassCodes C3360L; C3250; C3390 Article Type Applications; Practical Language English RecordType Conference ControlNo. 4687369 AbstractNos. C9407-3360L-041 ISBN or SBN 0 8186 3450 2 References 9 U.S. Copyright Clearance Center Code 0 8186 3450 2/93/$3.00 Country Pub. USA date 1214 ------------------------------------------------------------ Author Brown, G.; Forte, P.; Malyan, R.; Barnwell, P.; Mercury Commun., UK Title Perceptual feature based object recognition for automatic inspection Source Machine Vision Applications, Architectures, and Systems Integration II; Part: Boston, MA, USA; Part: 7-9 Sept. 1993; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2064; 1993; pp. 307-18 Abstract This paper describes the development of a practical object recognition framework for industrial inspection applications. The framework is being used to develop a prototype automatic inspection system for surface mounted electronic assemblies. The recognition technique uses the object oriented paradigm to both process and model the image information. Three inter-linked information structures, that allow both image and object modelling information to be represented at multiple resolutions, are described. The first structure is a hierarchical feature based description constructed from the recursive grouping of perceptual features. The second structure provides a containment tree to describe the connections between sub-component parts of a complex object. The final structure is the inheritance tree that provides the capability to model multiple variations of the same object type. The inspection module begins by processing image information from a newly developed nonlinear shape abstraction technique. The information is delivered in the form of a multi- resolution scale space description of the captured image Thesaurus assembling; automatic optical inspection; computer vision; feature extraction; printed circuit manufacture; surface mount technology Other Terms perceptual feature based object recognition; automatic inspection ; industrial inspection; surface mounted electronic assemblies; interlinked information structures; object modelling; hierarchical feature based description; recursive grouping; containment tree; image processing; inspection module; nonlinear shape abstraction; multiresolution scale space description ClassCodes B0170L; B2210D; B6140C; C5260B; C1250; C3355F Article Type Practical Coden PSISDG Language English RecordType Conference ControlNo. 4670634 AbstractNos. B9406-0170L-019; C9406-5260B-158 ISSN 0277786X References 12 U.S. Copyright Clearance Center Code 0 8194 1329 1/93/$6.00 Country Pub. USA date 1218 ------------------------------------------------------------ ------------------------------------------------------------ Author Bichsel, M.; Pentland, A.P.; Media Lab., MIT, Cambridge, MA, USA Title Human face recognition and the face image set's topology Source CVGIP: Image Understanding; CVGIP, Image Underst. (USA); vol.59, no.2; March 1994; pp. 254-61 Abstract Considering an n*n image as an n/sup 2/-dimensional vector, then images of faces can be considered as points in this n/sup 2/- dimensional image space. Previous studies of physical transformations of the face, including translation, small rotations, and illumination changes, showed that the set of face images consists of relatively simple connected subregions in image space. Consequently linear matching techniques can be used to obtain reliable face recognition. However, for more general transformations, such as large rotations or scale changes, the face subregions become highly non-convex. We have therefore developed a scale-space. matching technique that allows us to take advantage of knowledge about important geometrical transformations and about the topology of the face subregion in image space. While recognition of faces is the focus of the paper, the algorithm is sufficiently general to be applicable to a large variety of object recognition tasks Thesaurus face recognition; topology Other Terms face recognition; face image; geometrical transformations; topology; object recognition; physical transformations; small rotations ClassCodes C5260B; C1250; C1160 Article Type Theoretical / Mathematical Coden CIUNEJ Language English RecordType Journal ControlNo. 4668583 AbstractNos. C9406-5260B-137 ISSN 10499660 References 25 U.S. Copyright Clearance Center Code 1049-9660/94/$6.00 Country Pub. USA date 1225 ------------------------------------------------------------ Author Goshtasby, A.; Dept. of Electr. Eng. & Comput. Sci., Illinois Univ., Chicago, IL, USA Title On edge focusing (scale-space imaging) Source Image and Vision Computing; Image Vis. Comput. (UK); vol.12, no.4; May 1994; pp. 247-56 Abstract An algorithm is described that can track edges determined by the Laplacian of Gaussian operator from low to high resolution. The algorithm does not require a precomputed scale step size, but rather determines the step size adaptively by using the image content. A binary search is carried out to find the topology of the scale-space image, which is then used to track the edges. The proposed algorithm guarantees correct edge tracking with step sizes considerably larger than has been previously used Thesaurus edge detection Other Terms edge focusing; Gaussian operator; image content; binary search; scale-space image ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical Coden IVCODK Language English RecordType Journal ControlNo. 4666552 AbstractNos. B9406-6140C-157; C9406-1250-111 ISSN 02628856 References 16 U.S. Copyright Clearance Center Code 0262-8856/94/040247-10$10.00 Country Pub. UK date 1227 ------------------------------------------------------------ Author Ray, B.K.; Ray, K.S.; Electron. & Commun. Sci. Unit, Indian Stat. Inst., Calcutta, India Title A new approach to scale-space image Source Pattern Recognition Letters; Pattern Recognit. Lett. (Netherlands); vol.15, no.4; April 1994; pp. 365-72 Abstract This paper presents a new approach to scale-space image based on the authors' (1991) work on polygonal approximation. The polygonal approximation involves a counter which can be varied to produce different sets of vertex points and hence different polygonal approximations. These polygons are combined to construct a scale-space image. The scale-space image is in conformity with the scale-space theory. The scale-space behaviour of different corner models is also analysed and a method for corner detection is proposed. The validity of the proposed technique is tested through experimental results Thesaurus edge detection; finite difference methods; image processing; statistical analysis Other Terms scale-space image; polygonal approximation; vertex points; corner models; corner detection ClassCodes B6140C; B0290P; B0240Z; C1250; C4170; C1140Z Article Type Theoretical / Mathematical Coden PRLEDG Language English RecordType Journal ControlNo. 4666484 AbstractNos. B9406-6140C-151; C9406-1250-106 ISSN 01678655 References 9 U.S. Copyright Clearance Center Code 0167-8655/94/$07.00 Country Pub. Netherlands date 1226 ------------------------------------------------------------ Author Ter Haar Romeny, B.M.; Florack, L.M.J.; Salden, A.H.; Viergever, M.A.; 3D Computer Vision Res. Group, Utrecht Univ. Hospital, Netherlands Title Higher order differential structure of images Source Proceedings of 13th International Conference on Information Processing in Medical Imaging; Part: Flagstaff, AZ, USA; Part: 14-18 June 1993; Information Processing in Medical Imaging. 13th International Conference, IPMI '93 Proceedings; Berlin, Germany; Springer-Verlag; xvi+567; 1993; pp. 77-93 Editor Barrett, H.H.; Gmitro, A.F. Abstract Presents a tutorial on the basic concepts for vision in the 'Koenderink' school. The concept of scale-space is a necessity if the extraction of structure from measured physical signals (i.e. images) is being done. The Gaussian derivative kernels then are, for physical signals, the natural analogs of the mathematical differential operators. This paper discusses some interesting properties of the Gaussian derivative kernels, like their orthogonality and behaviour with noisy input data. Geometrical structure to be extracted is expressed as differential invariants, in this paper limited to invariants under orthogonal transformations. Three representations are summarized: Cartesian, gauge and manifest invariant notation. Many explicit examples are given. A section is included about computer implementation of the calculation of higher order invariant structure Thesaurus computer vision; geometry; image processing Other Terms higher order differential image structure; computer vision; geometrical structure extraction; Cartesian representation; gauge representation; scale-space; Gaussian derivative kernels; orthogonality; noisy input data; manifest invariant notation; higher order invariant structure ClassCodes B6140C; C1250; C5260B Article Type General or Review; Practical Language English RecordType Conference ControlNo. 4648047 AbstractNos. B9405-6140C-252; C9405-1250-180 ISBN or SBN 3 540 56800 X References 35 Country Pub. Germany date 1215 ------------------------------------------------------------ Author Brown, G.; Forte, P.; Malyan, R.; Mercury Commun. Ltd., London, UK Title Non-linear shape abstraction for automatic inspection Source Vision Geometry II; Part: Boston, MA, USA; Part: 9-10 Sept. 1993 ; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2060; 1993; pp. 125-32 Abstract The authors describe the implementation of a non-linear 2D shape abstraction technique. The abstraction procedure systematically simplifies the shape's description using a group of predefined 'rewrite rules'. This procedure operates on a new compact and efficient two dimensional shape representation. The abstraction technique generates a detailed scale space representation of the shape being abstracted. The authors provide a method for analysing this representation to determine the significant shape descriptions. These significant descriptions are used to produce a reduced scale space description of the shape, for use by an object recognition module. The work described will form a module in a system being developed at Kingston University, to inspect surface mounted technology (SMT) circuit boards for defects. These defects will include poor solder joint quality, component placement faults, missing components, etc. The inspection system uses images, captured from a CCD camera, which are initially processed using a transputer. Early test results are presented to show the benefits of this technique in comparison with Gaussian smoothing Thesaurus circuit analysis computing; computer vision; edge detection; inspection; surface mount technology Other Terms automatic inspection; non-linear 2D shape abstraction; abstraction procedure; predefined rewrite rules; two dimensional shape representation; scale space representation; reduced scale space description; object recognition module; surface mounted technology circuit boards; solder joint quality; component placement faults; CCD camera; transputer ClassCodes B1130B; C7410D; C5260B Article Type Practical Coden PSISDG Language English RecordType Conference ControlNo. 4626930 AbstractNos. B9405-1130B-001; C9405-7410D-006 ISSN 0277786X References 7 U.S. Copyright Clearance Center Code 0 8194 1325 9/93/$6.00 Country Pub. USA date 1218 ------------------------------------------------------------ Author Theimer, W.M.; Mallot, H.A.; Tolg, S.; Inst. fur Neuroinf., Ruhr-Univ., Bochum, Germany Title Phase method for binocular vergence control and depth reconstruction Source Intelligent Robots and Computer Vision XI: Biological, Neural Net and 3-D Methods; Part: Boston, MA, USA; Part: 18-20 Nov. 1992; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1826; 1992; pp. 76-87 Abstract We present a technique to guide vergence movements for an active stereo camera system and to construct dense disparity maps. Both processes are described in the same theoretical framework based on phase differences in complex Gabor filter responses, modelling receptive field properties in the visual cortex. While the camera movements are computed with coarse spatial resolution input images, disparity calculation uses finer resolutions in a scale space. The correspondence problem is solved implicitly by restricting the disparity range around zero disparity to the filter kernel sizes (Panum's area in the human visual system). The method contrasts to matching algorithms-that require an explicit search for correspondence-and to correlation, needing a maximum detection in the correlation function. The vergence process is interpreted as a mechanism to minimize global disparity, thereby setting a 3D region of interest for subsequent disparity detection. This small volume is centered around the fixation point where both optical axes intersect. Additionally it produces a scalar distance measure via vergence angles and camera base. The disparity map represents smaller local disparities as an important cue for depth perception. The vergence control works in a real-time feedback loop. Quantitative results are presented Thesaurus feedback; filtering and prediction theory; stereo image processing Other Terms phase method; binocular vergence control; depth reconstruction; active stereo camera system; complex Gabor filter responses; visual cortex; coarse spatial resolution input images; filter kernel sizes; Panum's area; global disparity minimization; fixation point; depth perception; real-time feedback loop ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4626626 AbstractNos. B9405-6140C-007; C9405-1250-008 ISSN 0277786X References 13 U.S. Copyright Clearance Center Code 0 8194 1027 6/92/$4.00 Country Pub. USA date 1207 ------------------------------------------------------------ Author Chaney, R.D.; Artificial Intelligence Lab., MIT, Cambridge, MA, USA Title Computation of the medial axis skeleton at multiple complexities Source Intelligent Robots and Computer Vision XI: Algorithms, Techniques and Active Vision; Part: Boston, MA, USA; Part: 16-18 Nov. 1992; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1825; 1992; pp. 469-80 Abstract The medial axis skeleton is a thin line graph that preserves the topology of a simply connected region. The skeleton has often been cited as a useful representation for shape description, region interpretation, and object recognition. Unfortunately, the computation of the skeleton is extremely sensitive to variations in the bounding contour. Tiny perturbations in the contour often lead to spurious branches of the skeleton. In this paper, we consider a robust method for computing the medial axis skeleton across a variety of scales. The scale-space is parametric with the complexity of the bounding contour. The complexity is defined as the number of extrema of curvature in the contour. A set of curves is computed to represent the bounding contour across a variety of complexity measures. The curves possessing larger complexity measures represent greater detail than curves with smaller measures. A medial axis skeleton is computed directly from each contour. The result is a set of skeletons that represent only the gross structure of the region at coarse scales (low complexity), but represent more of the detail at fine scales (high complexity) Thesaurus computational complexity; image processing; image recognition Other Terms medial axis skeleton; multiple complexities; thin line graph; topology; simply connected region; shape description; region interpretation; object recognition; bounding contour variations; scale-space; bounding contour ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4619983 AbstractNos. B9404-6140C-294; C9404-1250-184 ISSN 0277786X References 11 U.S. Copyright Clearance Center Code 0 8194 1026 8/92/$4.00 Country Pub. USA date 1207 ------------------------------------------------------------ Author Johansen, P.; Dept. of Comput. Sci., Copenhagen Univ., Denmark Title On the classification of toppoints in scale space Source Journal of Mathematical Imaging and Vision; J. Math. Imaging Vis. (Netherlands); vol.4, no.1; Jan. 1994; pp. 57-67 Abstract An algebraic classification scheme for toppoints in scale space is proposed. A critical point is a point whose spatial derivatives are zero, and a toppoint is a critical point in which the Hessian does not have full rank. A critical curve is a curve consisting of critical points. It is proposed that toppoints be classified according to the number of critical curves that intersect at the toppoint. Toppoints are analyzed further when one or two critical curves pass through the toppoint. Results on extrema, saddle points and intersections are found Thesaurus algebra; image recognition Other Terms toppoint classification; scale space; algebraic classification scheme; critical curves; extrema; saddle points; intersections ; image analysis ClassCodes B6140C; B0210; C1250; C1110 Article Type Theoretical / Mathematical Coden JIMVEC Language English RecordType Journal ControlNo. 4616880 AbstractNos. B9404-6140C-247; C9404-1250-149 ISSN 09249907 References 12 U.S. Copyright Clearance Center Code 0924-9907/94/$5.00 Country Pub. Netherlands date 1223 ------------------------------------------------------------ Author Kelch, J.; Wein, B.; Rogowski-Inst. for Electr. Eng., RWTH Aachen, Germany Title Model based segmentation of the tongue surface using a modified scale space filter Source Visual Communications and Image Processing '93; Part: Cambridge, MA, USA; Part: 8-11 Nov. 1993; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2094, pt.1; 1993; pp. 24-30 Abstract This edge detector is based on coarse-to-fine tracking by varying the smoothing parameter of the Laplacian-of-Gaussian filter (LoG). In this way contour segments of the tongue dorsum and other objects are extracted. A model supports identification of the tongue segments and interpolation of the surface in the spatiotemporal space. The tongue is formed as a chain of elliptical structure elements. This model stresses a direction to detect the orientation of the tongue and is flexible enough to form any shape. These structure elements are matched to the scale segments by correlation. A trainable cost path classifier selects the topological connections of the structure elements, which are linked by a spline interpolation. Virtual three-dimensional views of the contour surface in the spatiotemporal space are generated with different azimuthal angles for visualization Thesaurus biomedical ultrasonics; correlation methods; edge detection; image segmentation; medical image processing; physiological models; spatial filters; splines (mathematics); topology; tracking Other Terms model-based segmentation; ultrasonic imaging; modified scale space filter; edge detector; coarse-to-fine tracking; tongue segments; elliptical structure elements; trainable cost path classifier; topological connections; spline interpolation; visualization ClassCodes A8760B; A4230V; A0260; B7510B; B7820; B6140C; B0290F; C7330 ; C5260B; C4130 Article Type Applications; Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4614536 AbstractNos. A9408-8760B-003; B9404-7510B-029; C9404-7330-074 ISSN 0277786X References 11 U.S. Copyright Clearance Center Code 0 8194 1369 0/93/$6.00 Country Pub. USA date 1220 ------------------------------------------------------------ ------------------------------------------------------------ Author Alvarez, L.; Guichard, F.; Lions, P.-L.; Morel, J.-M.; Dept. de Inf. y Sistemas, Las Palmas Univ., Spain Title Axioms and fundamental equations of image processing Source Archive for Rational Mechanics and Analysis; Arch. Ration. Mech. Anal. (Germany); vol.123, no.3; 1993; pp. 199-257 Abstract Image-processing transforms must satisfy a list of formal requirements. We discuss these requirements and classify them into three categories: 'architectural requirements' like locality, recursivity and causality in the scale space, 'stability requirements' like the comparison principle and 'morphological requirements', which correspond to shape-preserving properties (rotation invariance, scale invariance, etc.). A complete classification is given of all image multiscale transforms satisfying these requirements. This classification yields a characterization of all classical models and includes new ones, which all are partial differential equations. The new models we introduce have more invariance properties than all the previously known models and in particular have a projection invariance essential for shape recognition. Numerical experiments are presented and compared. The same method is applied to the multiscale analysis of movies. By introducing a property of Galilean invariance, we find a single multiscale morphological model for movie analysis Thesaurus image processing; partial differential equations; transforms Other Terms image processing; transforms; architectural requirements; locality; recursivity; causality; stability requirements; comparison principle; morphological requirements; shape- preserving properties; rotation invariance; scale invariance; image multiscale transforms; partial differential equations; movies; Galilean invariance; multiscale morphological model ClassCodes B6140C; B0230; B0220; C1250; C4170 Article Type Theoretical / Mathematical Coden AVRMAW Language English RecordType Journal ControlNo. 4603142 AbstractNos. B9404-6140C-001; C9404-1250-001 ISSN 00039527 References 82 Country Pub. Germany date 1210 ------------------------------------------------------------ Author Ottenberg, K.; Neumann, H.; Stiehl, H.S.; Philips GmbH Forschungslab., Hamburg, Germany Title Quantitative description and reconstruction of intensity functions using scale-space and multiresolution processing Source Signal Processing VI - Theories and Applications. Proceedings of EUSIPCO-92, Sixth European Signal Processing Conference; Part: Brussels, Belgium; Part: 24-27 Aug. 1992; Sponsored by: Belgian Nat. Fund for Sci. Res.; CERA; LMS Int; Amsterdam, Netherlands; Elsevier; 3 vol. lvii+1844; 1992; pp. 1425-8 vol. 3 Editor Vandewalle, J.; Boite, R.; Moonen, M.; Oosterlinck, A. Abstract Differently scaled 1-D intensity discontinuities along the gradient direction define significant image structure intrinsically related to different physical causes in the scene, e.g. step-edges in images due to cast shadow boundaries or gradually changing edge profiles due to self shadow boundaries on smoothly curved object surfaces. Reliable detection and quantitative description of such a variety of scaled intensity discontinuities (with a priori unknown relative contrasts, widths, loci and orientations) in images require a family of appropriately scaled operators and a generally applicable quantitative multiscale processing scheme. Hence the authors propose (i) a multiscale scheme following Marr's (1976) and Korn's (1988) approach (ii) briefly present the theoretical framework for the 1-D case, and (iii) briefly report on numerical results from initial experiments with discrete 1-D data Thesaurus image processing Other Terms scale-space processing; image structure description; intensity functions; multiresolution processing; 1-D intensity discontinuities; gradient direction; image structure; edge profiles; shadow boundaries; relative contrasts; widths; loci; orientations; scaled operators; multiscale processing ClassCodes B6140C Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4584721 AbstractNos. B9403-6140C-084 ISBN or SBN 0 444 89587 6 References 14 Country Pub. Netherlands date 1204 ------------------------------------------------------------ Author Laine, A.; Fan, J.; Dept. of Comput. & Inf. Sci., Florida Univ., Gainesville, FL, USA Title Texture classification by wavelet packet signatures Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.15, no.11; Nov. 1993; pp. 1186-91 Abstract This correspondence introduces a new approach to characterize textures at multiple scales. The performance of wavelet packet spaces are measured in terms of sensitivity and selectivity for the classification of twenty-five natural textures. Both energy and entropy metrics were computed for each wavelet packet and incorporated into distinct scale space representations, where each wavelet packet (channel) reflected a specific scale and orientation sensitivity. Wavelet packet representations for twenty-five natural textures were classified without error by a simple two-layer network classifier. An analyzing function of large regularity (D/sub 20/) was shown to be slightly more efficient in representation and discrimination than a similar function with fewer vanishing moments (D/sub 6/) In addition, energy representations computed from the standard wavelet decomposition alone (17 features) provided classification without error for the twenty-five textures included in our study. The reliability exhibited by texture signatures based on wavelet packets analysis suggest that the multiresolution properties of such transforms are beneficial for accomplishing segmentation, classification and subtle discrimination of texture Thesaurus feature extraction; feedforward neural nets; image recognition; wavelet transforms Other Terms texture classification; wavelet packet signatures; scale- independence; wavelet packet spaces; sensitivity; selectivity; energy metrics; entropy metrics; scale space representations; scale sensitivity; orientation sensitivity; two-layer network classifier ClassCodes B6140C; C1250; C1230D Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4580750 AbstractNos. B9403-6140C-039; C9403-1250-028 ISSN 01628828 References 40 U.S. Copyright Clearance Center Code 0162-8828/93/$03.00 Country Pub. USA date 1220 ------------------------------------------------------------ Author Eggert, D.W.; Bowyer, K.W.; Dyer, C.R.; Christensen, H.I.; Goldgof, D.B.; Dept. of Comput. Sci. & Eng., Univ. of South Florida, Tampa, FL, USA Title The scale space aspect graph Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.15, no.11; Nov. 1993; pp. 1114-30 Abstract Currently the aspect graph is computed from the theoretical standpoint of perfect resolution in object shape, the viewpoint and the projected image. This means that the aspect graph may include details that an observer could never see in practice. Introducing the notion of scale into the aspect graph framework provides a mechanism for selecting a level of detail that is "large enough" to merit explicit representation. This effectively allows control over the number of nodes retained in the aspect graph. This paper introduces the concept of the scale space aspect graph, defines three different interpretations of the scale dimension, and presents a detailed example for a simple class of objects, with scale defined in terms of the spatial extent of features in the image Thesaurus graph theory; image processing Other Terms scale space aspect graph; projected image; scale dimension; image processing ClassCodes B6140C; B0250; C1250; C1160 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4580745 AbstractNos. B9403-6140C-036; C9403-1250-025 ISSN 01628828 References 39 U.S. Copyright Clearance Center Code 0162-8828/93/$03.00 Country Pub. USA date 1220 ------------------------------------------------------------ Author Lindeberg, T.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title Effective scale: a natural unit for measuring scale-space lifetime Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.15, no.10; Oct. 1993; pp. 1068-74 Abstract A manner in which a notion of effective scale can be introduced in a formal way is developed. For continuous signals, a scaling argument directly gives a natural unit for measuring scale-space lifetime in terms of the logarithm of the ordinary scale parameter. That approach is, however, not appropriate for discrete signals since an infinite lifetime would be assigned to structures existing in the original signal. It is shown how such an effective scale parameter can be defined to give consistent results for both discrete and continuous signals. The treatment is based on the assumption that the probability that a local extremum disappears during a short-scale interval should not vary with scale. As a tool for the analysis, estimates are given of how the density of local extrema can be expected to vary with scale in the scale-space representation of different random noise signals both in the continuous and discrete cases Thesaurus image processing Other Terms effective scale; scale-space lifetime measurement unit ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4580739 AbstractNos. B9403-6140C-031; C9403-1250-020 ISSN 01628828 References 20 U.S. Copyright Clearance Center Code 0162-8828/93/$03.00 Country Pub. USA date 1219 ------------------------------------------------------------ Author Neumann, H.; Ottenberg, K.; Hamburg Univ., Germany Title Estimating ramp-edge attributes from scale-space Source Signal Processing VI - Theories and Applications. Proceedings of EUSIPCO-92, Sixth European Signal Processing Conference; Part: Brussels, Belgium; Part: 24-27 Aug. 1992; Sponsored by: Belgian Nat. Fund for Sci. Res.; CERA; LMS Int; Amsterdam, Netherlands; Elsevier; 3 vol. lvii+1844; 1992; pp. 603-6 vol.1 Editor Vandewalle, J.; Boite, R.; Moonen, M.; Oosterlinck, A. Abstract Ramp-edges are frequently employed for modelling transitions of the image intensity function between regions of constant intensity value. For the domain of magnetic resonance images (MRI), this type of 'discontinuity' model can explicitly be derived from the underlying scene and the characteristics of the measurement process (e.g. partial-volume effects). For an automatic analysis of MRI, an estimation of the attributes defining such a ramp-edge is essential. Other models for the smooth image-transitions between plateaus corresponding to different anatomical structures, generated by partial-volume effects, within MRI have been investigated in detail by the authors (1989, 1992). These models are based on step-edge models convolved with Gaussian kernels, i.e. error-functions of different variance. An analysis of the accuracy of regularized first- and second-order differential operators for the localization of discontinuities (including ramp-edges) has been published by the authors (1990). The authors describe a procedure to estimate the remaining attributes of a ramp-edge, the transition width and the local contrast, from the scale-space representation of the ramp-edge intensity function, generated by the Gaussian convolution kernels Thesaurus biomedical NMR; edge detection; medical image processing Other Terms ramp-edge attributes estimation; image intensity function; constant intensity value; magnetic resonance images; partial- volume effects; MRI; step-edge models; Gaussian kernels; error-functions; second-order differential operators; transition width; local contrast; scale-space representation ClassCodes A8760G; A8770E; B7510B; B6140C Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4576307 AbstractNos. A9404-8760G-029; B9402-7510B-137 ISBN or SBN 0 444 89587 6 References 0 Country Pub. Netherlands date 1204 ------------------------------------------------------------ Author Wong, Y.-F.; Dept. of Electr. Eng., California Inst. of Technol., Pasadena, CA, USA Title Clustering data in the scale space Source ICARCV '92. Second International Conference on Automation, Robotics and Computer Vision; Part: Singapore; Part: 16-18 Sept. 1992; Sponsored by: IEE; Inst. Meas.& Control; Econom. Development Board; et al; Singapore; Nanyang Technol. Univ; 3 vol. (viii+934+viii+861+vii+908); 1992; pp. CV-20.2/1-5 vol.1 Abstract We derive a scale-space clustering algorithm based on information theory and statistical mechanics. We introduce the concept of cluster independence into clustering. The cluster centers correspond to the local minima of a thermodynamical free energy, which can be analyzed using nonlinear dynamics. The algorithm works by melting the system to produce a tree of clusters in the scale space. Melting can also handle variability in cluster densities, cluster sizes and ellipsoidal shapes. We tested successfully the algorithm on both simulated data and real data Thesaurus dynamics; information theory; pattern recognition; statistical mechanics Other Terms data analysis; scale-space clustering algorithm; information theory; statistical mechanics; local minima; thermodynamical free energy; nonlinear dynamics ClassCodes B6140C; B0240Z; B6110; C1250; C1140Z; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4575742 AbstractNos. B9402-6140C-283; C9402-1250-217 References 19 Country Pub. Singapore date 1205 ------------------------------------------------------------ Author Stach, J.; Shaw, S.; SRI Int., Menlo Park, CA, USA Title Robust relational trees by scale-space filtering Source Proceedings of the IEEE-SP International Symposium Time-Frequency and Time-Scale Analysis (Cat.No.92TH0478-8); Part: Victoria, BC, Canada; Part: 4-6 Oct. 1992; Sponsored by: IEEE; New York, NY, USA; IEEE; 577; 1992; pp. 173-6 Abstract A relational tree (RT) is a signal representation used to discriminate between multidimensional signals with arbitrary nonlinear monotonic distortion. The use of RTs is limited by noise. Scale-space filtering takes advantage of the inherently scale-like properties of an RT to provide a more robust signal representation across distortions. Some methods and properties of Gaussian scale-space filtering of RTs are examined. Scale trees (STs) and conventionally filtered (fixed-scale) RTs have been shown to be subsets of this process. Since the effect of scale- space filtering is to move segmentation uncertainty toward the leaves of the tree, other operations performed in the tree domain, such as filtering, can be optimized as well Thesaurus filtering and prediction theory; signal processing; trees (mathematics) Other Terms optimisation; Gaussian filtering; scale-space filtering; relational tree; signal representation; multidimensional signals ; nonlinear monotonic distortion; noise; inherently scale-like properties; segmentation uncertainty; leaves ClassCodes B6140; B0250; C1260; C1160 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4575484 AbstractNos. B9402-6140-164; C9402-1260-127 ISBN or SBN 0 7803 0805 0 References 5 U.S. Copyright Clearance Center Code 0 7803 0805 0/92/$3.00 Country Pub. USA date 1206 ------------------------------------------------------------ Author Lindeberg, T.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title Discrete derivative approximations with scale-space properties: a basis for low-level feature extraction Source Journal of Mathematical Imaging and Vision; J. Math. Imaging Vis. (Netherlands); vol.3, no.4; Nov. 1993; pp. 349-76 Abstract This article shows how discrete derivative approximations can be defined so that scale-space properties hold exactly also in the discrete domain. An axiomatic derivation of how a multiscale representation of derivative approximations can be constructed from a discrete signal, so that it possesses an algebraic structure similar to that possessed by the derivatives of the traditional scale-space representation in the continuous domain, is given. A family of kernels is derived, that constitutes discrete analogues to the continuous Gaussian derivatives. The representation has theoretical advantages over other discretizations of the scale-space theory in the sense that the operators that commute before discretization also commute after discretization. Some computational implications of this are that derivative approximations can be computed directly from smoothed data and that this will give exactly the same result as convolution with the corresponding derivative approximation kernel. Moreover, a number of normalization conditions are automatically satisfied. The proposed methodology leads to a scheme of computations of multiscale low-level feature extraction that is conceptually very simple Thesaurus computer vision; edge detection; feature extraction; filtering and prediction theory; image recognition Other Terms discrete derivative approximations; scale-space; low-level feature extraction; axiomatic derivation; multiscale representation; discrete signal; algebraic structure; discretization; convolution; normalization conditions; edge detection; compute vision ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Coden JIMVEC Language English RecordType Journal ControlNo. 4557269 AbstractNos. B9402-6140C-024; C9402-1250-013 ISSN 09249907 References 36 U.S. Copyright Clearance Center Code 0924-9907/93/$5.00 Country Pub. Netherlands date 1220 ------------------------------------------------------------ Author Florack, L.M.J.; Ter Haar Romeny, B.M.; Koenderink, J.J.; Viergever, M.A.; Utrecht Univ., Netherlands Title Cartesian differential invariants in scale-space Source Journal of Mathematical Imaging and Vision; J. Math. Imaging Vis. (Netherlands); vol.3, no.4; Nov. 1993; pp. 327-48 Abstract We present a formalism for studying local image structure in a systematic, coordinate-independent, and robust way, based on scale-space theory, tensor calculus, and the theory of invariants. We concentrate on differential invariants. The formalism is of general applicability to the analysis of grey-tone images of various modalities, defined on a D-dimensional spatial domain. We propose a "diagrammar" of differential invariants and tensors, i. e., a diagrammatic representation of image derivatives in scale- space together with a set of simple rules for representing meaningful local image properties. All local image properties on a given level of inner scale can be represented in terms of such diagrams, and, vice versa, all diagrams represent coordinate- independent combinations of image derivatives, i.e., true image properties. We present complete and irreducible sets of (nonpolynomial) differential invariants appropriate for the description of local image structure up to any desired order. Any differential invariant can be expressed in terms of polynomial invariants, pictorially represented by closed diagrams. Here we consider a complete, irreducible set of polynomial invariants up to second order (inclusive) Thesaurus image processing; invariance; tensors Other Terms Cartesian differential invariants; scale-space; local image structure; tensor calculus; grey-tone images; D-dimensional spatial domain; diagrammar; image derivatives; true image properties; polynomial invariants ClassCodes B6140C; B0210; C1250; C1110 Article Type Theoretical / Mathematical Coden JIMVEC Language English RecordType Journal ControlNo. 4557268 AbstractNos. B9402-6140C-023; C9402-1250-012 ISSN 09249907 References 48 U.S. Copyright Clearance Center Code 0924-9907/93/$5.00 Country Pub. Netherlands date 1220 ------------------------------------------------------------ Author Snyder, W.E.; Youn-Sik Han; Bilbro, G.L.; Dept. of Radiol., Med. Center Boulevard, Winston-Salem, NC, USA Title A unified theory of edge-preserving smoothing Source Artificial Neural Networks, 2. Proceedings of the 1992 International Conference (ICANN-92); Part: Brighton, UK; Part: 4-7 Sept. 1992; Sponsored by: UK DTI; Eur. Commission; Amsterdam, Netherlands; Elsevier; 2 vol. (xviii+xxx+1700); 1992; pp. 1675-83 vol.2 Editor Aleksander, I. Abstract Two edge-preserving smoothing techniques are discussed, mean field annealing, graduated non-convexity, and compared to a feature extraction technique known as variable conductance diffusion. In previous literature, the first two techniques have been shown to be equivalent. The third technique is shown to also be equivalent. Furthermore, operations across scale space are shown to be equivalent to annealing. The demonstration of the mathematical equivalence of three independently derived and successful methods leads to conclusions concerning the fundamental nature of image analysis algorithms Thesaurus feature extraction; image processing Other Terms unified theory; edge-preserving smoothing techniques; mean field annealing; graduated non-convexity; feature extraction technique; variable conductance diffusion; mathematical equivalence; image analysis algorithms ClassCodes C5260B Article Type Practical Language English RecordType Conference ControlNo. 4553441 AbstractNos. C9401-5260B-232 ISBN or SBN 0 444 89488 8 References 26 Country Pub. Netherlands date 1205 ------------------------------------------------------------ Author Sapiro, G.; Tannenbaum, A.; Technion-Israel Inst. of Technol., Haifa, Israel Title Affine invariant scale-space Source International Journal of Computer Vision; Int. J. Comput. Vis. (Netherlands); vol.11, no.1; Aug. 1993; pp. 25-44 Abstract A new affine invariant scale-space for planar curves is presented. The scale-space is obtained from the solution of a novel nonlinear curve evolution equation which admits affine invariant solutions. This flow was proved to be the affine analogue of the well known Euclidean shortening flow. The evolution also satisfies properties such as causality, which makes it useful in defining a scale-space. Using an efficient numerical algorithm for curve evolution, this continuous affine flow is implemented, and examples are presented. The affine-invariant progressive smoothing property of the evolution equation is demonstrated well Thesaurus computational geometry Other Terms affine invariant scale-space; planar curves; Euclidean shortening flow; causality; progressive smoothing property ClassCodes C4260 Article Type Bibliography/Literature Suvery; Practical; Theoretical / Mathematical Coden IJCVEQ Language English RecordType Journal ControlNo. 4550660 AbstractNos. C9401-4260-066 ISSN 09205691 References 72 Country Pub. Netherlands date 1217 ------------------------------------------------------------ Author Lindberg, T.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title Detecting salient blob-like image structures and their scales with a scale-space primal sketch: a method for focus-of-attention Source International Journal of Computer Vision; Int. J. Comput. Vis. (Netherlands); vol.11, no.3; Dec. 1993; pp. 283-318 Abstract Presents: (i) a multiscale representation of grey-level shape called the scale-space primal sketch, which makes explicit both features in scale-space and the relations between structures at different scales, (ii) a methodology for extracting significant blob-like image structures from this representation, and (iii) applications to edge detection, histogram analysis, and junction classification demonstrating how the proposed method can be used for guiding later-stage visual processes. The representation gives a qualitative description of image structure, which allows for detection of stable scales and associated regions of interest in a solely bottom-up data-driven way. In other words, it generates coarse segmentation cues, and can hence be seen as preceding further processing, which can then be properly tuned. It is argued that once such information is available, many other processing tasks can become much simpler. Experiments on real imagery demonstrate that the proposed theory gives intuitive results Thesaurus computer vision; edge detection Other Terms salient; scale-space primal sketch; multiscale representation; grey-level shape; edge detection; histogram analysis; junction classification; image structure; coarse segmentation; blob-like ClassCodes C5260B; B6140C; C1250 Article Type Bibliography/Literature Suvery; Practical; Theoretical / Mathematical Coden IJCVEQ Language English RecordType Journal ControlNo. 4544297 AbstractNos. B9401-6140C-236; C9401-5260B-142 ISSN 09205691 References 72 U.S. Copyright Clearance Center Code 0920-5691/93/$5.00 Country Pub. Netherlands date 1221 ------------------------------------------------------------ Author Taylor, J.R.; Olson, T.J.; Dept. of Comput. Sci., Virginia Univ., Charlottesville, VA, USA Title Precise vergence control in complex scenes Source Proceedings of the SPIE - The International Society for Optical Engineering; Proc. SPIE - Int. Soc. Opt. Eng. (USA); vol.2056; 1993; pp. 22-30 Abstract In binocular systems, vergence is the process of directing the gaze so that the optical axes intersect at the point of interest. Region based methods of disparity analysis provide fast and reliable estimates of the vergence error, but it is difficult to determine on what image features these approaches are in fact verging. Previous approaches to vergence control have for the most part failed to ensure that both cameras actually verge on the object of interest, especially in complex scenes. This paper presents a system that addresses this problem. By using the cepstral filter in a multiresolution setting with a dominant camera, the system can verge accurately in complex scenes. Specifically, the system adaptively refines the vergence angle in a scale space consisting of the center patches on a Gaussian pyramid. The effects of the cepstrum in a multiresolution system are analyzed, and the precision and performance of the new system are verified on natural scenes Thesaurus cameras; image processing; optical filters; optical information processing; spectral analysis Other Terms image processing; vergence control; complex scenes; optical axes; cameras; cepstral filter; vergence angle; Gaussian pyramid; cepstrum; multiresolution system ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Journal ControlNo. 4542570 AbstractNos. B9401-6140C-197; C9401-1250-125 ISSN 0277786X References 12 U.S. Copyright Clearance Center Code 0 8194 1321 6/93/$6.00 Country Pub. USA date 1210 ------------------------------------------------------------ Author Djemel Ziou; Tabbone, S.; EERIE-LERI, Nimes, France Title A multi-scale edge detector Source Pattern Recognition; Pattern Recognit. (UK); vol.26, no.9; Sept. 1993; pp. 1305-14 Abstract A multi-scale edge detector with subpixel accuracy is described. A subpixel Laplacian edge detector, recursively implemented, is run at different scales and the recovered edge information is combined. The multi-scale edge detection is based on the behavior of edges in scale space and takes into account their physical phenomena. With this purpose in mind, four step edge models are considered: the ideal, the blurred, the pulse and the staircase. It is emphasized that the use of two scales (the larger and the smaller) is sufficient for good edge detection. Furthermore, a set of rules is derived for combining edge information obtained from a Laplacian detector which has some special properties. However, this type of edge detector gives at least two classes of false edges, one of which cannot be eliminated by the usual thresholding methods. An appropriate thresholding algorithm is given taking into account the origin of the false edges and their behavior in scale space Thesaurus edge detection Other Terms multi-scale edge detector; subpixel accuracy; subpixel Laplacian edge detector; scale space; step edge models; thresholding algorithm; ideal model; blurred model; pulse model ; staircase model ClassCodes C1250; B6140C; C5260B Article Type Practical; Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 4528349 AbstractNos. B9401-6140C-048; C9401-1250-029 ISSN 00313203 References 24 U.S. Copyright Clearance Center Code 0031-3203/93/$6.00+.00 Country Pub. UK date 1218 ------------------------------------------------------------ Author Barth, E.; Zetzsche, C.; Ferraro, M.; Rentschler, I.; Inst. fur Medizinische Psychol., Munchen, Germany Title Fractal properties from 2D-curvature on multiple scales Source Geometric Methods in Computer Vision II; Part: San Diego, CA, USA ; Part: 12-13 July 1993; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2031; 1993; pp. 87-99 Abstract Basic properties of 2D-nonlinear scale-space representations of images are considered. Local-energy filters are used to estimate the Hausdorff dimension, D/sub H/, of images. A new fractal dimension, D/sub N/, defined as a property of 2D-curvature representations on multiple scales, is introduced as a natural extension of traditional fractal dimensions, and it is shown that the two types of fractal dimensions can give a less ambiguous description of fractal image structure. Some more general properties of curvature representations on multiple scales are considered. Simulations are used to analyse the stability of curvature maxima across scale and to illustrate that spurious resolution can be avoided by extraction of 2D-curvature features Thesaurus computer vision; fractals; image processing Other Terms Hausdorff dimension; 2D-curvature representations on multiple scales; fractal dimensions; fractal image structure; stability; spurious resolution ClassCodes B6140C; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4514973 AbstractNos. B9312-6140C-162; C9312-5260B-106 ISSN 0277786X References 25 U.S. Copyright Clearance Center Code 0 8194 1280 5/93/$6.00 Country Pub. USA date 1216 ------------------------------------------------------------ Author Salden, A.H.; ter Haar Romeny, B.M.; Viergever, M.A.; 3D Comput. Vision Res. Group, Utrecht Univ. Hospital, Netherlands Title Affine and projective differential geometric invariants of space curves Source Geometric Methods in Computer Vision II; Part: San Diego, CA, USA ; Part: 12-13 July 1993; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2031; 1993; pp. 64-74 Abstract By means of classical scale space theory, algebraic invariance theory and classical differential geometry a new method of shape description for space curves from one or multiple views is proposed in terms of complete and irreducible sets of invariants. The method is based on defining implicitly connections for the observed curves that are highly correlated to the projected space curves, assumed to reveal themselves as coherent structures in the scale space representation of the differential structure of the input images. Several applications to stereo, optic flow, texture analysis and image matching are indicated briefly Thesaurus computer vision; curve fitting; differential geometry; image texture; stereo image processing; surface fitting Other Terms affine differential geometric invariants; projective differential geometric invariants; scale space theory; algebraic invariance theory; shape description; space curves; coherent structures; stereo; optic flow; texture analysis; image matching ClassCodes B6140C; B0250; C5260B; C1160; C4260 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4514971 AbstractNos. B9312-6140C-160; C9312-5260B-104 ISSN 0277786X References 28 U.S. Copyright Clearance Center Code 0 8194 1280 5/93/$6.00 Country Pub. USA date 1216 ------------------------------------------------------------ Author Chen, J.G.; Tian, Q.; Ren, X.H.; Comput. Center, Taiyuan Univ. of Technol., Shanxi, China Title Adaptive edge focusing Source Communications on the Move. Singapore. ICCS/ISITA '92(Cat. No. 92TH0479-6); Part: Singapore; Part: 16-20 Nov. 1992; Sponsored by: IEEE; Singapore Telecommn.; Telecommn.Authority Singapore; et al; New York, NY, USA; IEEE; 3 vol. (xxvii+1422); 1990; pp. 624-32 vol.2 Editor Ng, C.S.; Yeo, T.S.; Yeo, S.P. Abstract Edge focusing, an efficient implementation method of continuous scale space filtering, has achieved better result in edge detection. Yet, the edge focusing algorithm smooths all edges with the filter of same scale, that generates ragged edges in the following areas: (1) area with diffuse edge, (2) texture edges. An optimal scale of edge detector is derived. With this optimal scale, the original edge focusing algorithm is modified. During the process of the revised edge focusing, if the scale of the edge detector has reached the optimal scale at any edge point, then the edge at that point will not be changed any more with a decreasing scale of the filter. The benefit of this modification is that it can avoid ragged edge and false edge points. The performance of the modified edge focusing algorithm is better than the original one. The experimental results support this conclusion Thesaurus edge detection Other Terms adaptive edge focusing; continuous scale space filtering; edge detection; edge focusing algorithm; diffuse edge; texture edges ; optimal scale ClassCodes B6140C; C1250 Article Type Experimental Language English RecordType Conference ControlNo. 4513283 AbstractNos. B9312-6140C-146; C9312-1250-116 ISBN or SBN 0 7803 0803 4 References 18 Country Pub. USA date 1207 ------------------------------------------------------------ Author Mori, K.; Matsuda, M.; Doi, S.; Takahashi, H.; Shimizu, E.; Dept. of Electron., Osaka Electro-Commun. Univ., Japan Title The real-time underwater data transmission system using scale- space filtering Source Parallel Computing and Transputer Applications; Part: Barcelona, Spain; Part: 21-25 Sept. 1992; Barcelona, Spain; CIMNE; 2 vol. 1520; 1992; pp. 356-65 vol.1; Available from: IOS Press, Amsterdam, Netherlands Editor Valero, M.; Onate, E.; Jane, M.; Larriba, J.L.; Suarez, B. Abstract The underwater data transmission using scale-space filtering and fingerprint pattern keying as the data transmission method has been proposed. The matching method using the number of the zero- cross points in fingerprint pattern keying is proposed as one of the data transmission methods which is conscious of the real-time processing. The real-time underwater data transmission system based on the parallel connected transputer array is designed Thesaurus real-time systems; signal processing; transputer systems; underwater sound Other Terms real-time underwater data transmission; scale-space filtering; fingerprint pattern keying; matching method; zero-cross points; parallel connected transputer array ClassCodes A4330; B6270; C5260; C5220P Article Type Practical Language English RecordType Conference ControlNo. 4508968 AbstractNos. A9323-4330-005; B9312-6270-001; C9312-5260-006 ISBN or SBN 84 87867 13 8 References 3 Country Pub. Spain date 1205 ------------------------------------------------------------ Author Campos, J.C.; Linney, A.D.; Moss, J.P.; Univ. Coll., London, UK Title The analysis of facial profiles using scale space techniques Source Pattern Recognition; Pattern Recognit. (UK); vol.26, no.6; June 1993; pp. 819-24 Abstract A method is presented of analysing facial profiles by using scale space techniques, reported in the pattern recognition literature. Through this technique landmarks, which are some extremal points on the profile are mathematically defined in order to divide the profile into a number of regions considered suitable for analysis. They are seen to correspond to parts of the face of interest to the clinician. Thus, the whole profile can be qualitatively described by the scale space image and the shape of individual regions quantitatively described by a curvature value. A mid-line facial profile is used to illustrate the method and a cleft palate case is analysed and the assessment of changes due to surgical correction are shown Thesaurus face recognition; image segmentation Other Terms image processing; profile segmentation; shape description; Gaussian smoothing; zero crossings; facial profiles; scale space techniques; pattern recognition; landmarks; curvature value ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 4507623 AbstractNos. B9312-6140C-039; C9312-1250-032 ISSN 00313203 References 17 U.S. Copyright Clearance Center Code 0031-3203/93/$6.00+.00 Country Pub. UK date 1215 ------------------------------------------------------------ Author Wong, Y.; Posner, E.C.; Lawrence Livermore Nat. Lab., CA, USA Title A new clustering algorithm applicable to multispectral and polarimetric SAR images Source IEEE Transactions on Geoscience and Remote Sensing; IEEE Trans. Geosci. Remote Sens. (USA); vol.31, no.3; May 1993; pp. 634-44 Abstract The authors applied a scale-space clustering algorithm to the classification of a multispectral and polarimetric SAR image of an agricultural site. After the initial polarimetric and radiometric calibration and noise cancellation, a 12-dimensional feature vector for each pixel was extracted from the scattering matrix. The clustering algorithm partitioned a set of unlabeled feature vectors from 13 selected sites, each site corresponding to a distinct crop, into 13 clusters without any supervision. The cluster parameters were then used to classify the whole image. The classification map is much less noisy and more accurate than those obtained by hierarchical rules. Starting with every point as a cluster, the algorithm works by melting the system to produce a tree of clusters in the scale space. It can cluster data in any multidimensional space and its insensitive to variability in cluster densities, sizes and ellipsoidal shapes. This algorithm, more powerful than existing ones, may be useful for remote sensing for land use Thesaurus image recognition; remote sensing by radar Other Terms multispectral SAR; synthetic aperture radar; clustering algorithm; polarimetric SAR images; classification; agricultural site; 12-dimensional feature vector; scattering matrix; unlabeled feature vectors; crop; melting; tree of clusters; scale space; multidimensional space; remote sensing; land use ClassCodes B7730; B6320; B6140C Article Type Applications; Theoretical / Mathematical Coden IGRSD2 Language English RecordType Journal ControlNo. 4505451 AbstractNos. B9312-7730-002 ISSN 01962892 References 26 U.S. Copyright Clearance Center Code 0196-2892/93/$03.00 Country Pub. USA date 1214 ------------------------------------------------------------ Author Roca, F.X.; Binefa, X.; Vitria, J.; Dept. d'Inf., Univ. Autonoma de Barcelona, Spain Title Multiscale structure extraction using morphological tools: applications to edge detection Source Image Algebra and Morphological Image Processing IV; Part: San Diego, CA, USA; Part: 12-13 July 1993; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2030; 1993; pp. 231-9 Abstract The purpose of multiscale analysis is to extract information from the original image features, studying their behaviour through various scale levels. This paper presents a new methodology, based on the use of tools provided by Mathematical Morphology, applied to scale-space image feature tracing. This methodology is here intended to solve two concrete problems: edge detection and depth perception Thesaurus edge detection; feature extraction; mathematical morphology Other Terms image processing; structure extraction; morphological tools; edge detection; multiscale analysis; methodology; scale-space image feature tracing; depth perception ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4504571 AbstractNos. B9312-6140C-022; C9312-1250-016 ISSN 0277786X References 17 U.S. Copyright Clearance Center Code 0 8194 1279 1/93/$6.00 Country Pub. USA date 1216 ------------------------------------------------------------ Author Jackway, P.T.; Centre for Signal Processing Res., Queensland Univ. of Technol., Brisbane, Qld., Australia Title Scale space properties of the multiscale morphological closing- opening filter Source ICIP 92. Proceedings of the 2nd Singapore International Conference on Image Processing; Part: Singapore; Part: 7-11 Sept. 1992; Singapore; World Scientific; xxii+734; 1992; pp. 278-81 Editor Srinivasa, V.; Ong Sim Heng; Ang Yew Hock Abstract Scale-space is an important concept used in image and signal processing and pattern recognition. Traditional scale-space is generated by a linear Gaussian smoothing operation. The author presents a nonlinear smoother corresponding to the multiscale opening and closing operations of mathematical morphology which also generates a scale-space. He shows that a parabolic structuring element possesses desirable properties and demonstrates the necessary monotonic scale-space property for this structuring element Thesaurus filtering and prediction theory; image processing; mathematical morphology; pattern recognition; signal processing Other Terms image processing; multiscale morphological closing-opening filter ; signal processing; pattern recognition; nonlinear smoother; mathematical morphology; parabolic structuring element; scale- space property ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4499038 AbstractNos. B9311-6140C-240; C9311-1250-171 ISBN or SBN 981 02 1182 1 References 11 Country Pub. Singapore date 1205 ------------------------------------------------------------ Author Kimia, B.B.; Tannenbaum, A.R.; Zucker, S.W.; Brown Univ., Providence, RI, USA Title Non-linear shape approximation via the entropy scale space Source Geometric Methods in Computer Vision II; Part: San Diego, CA, USA ; Part: 12-13 July 1993; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.2031; 1993; pp. 218-33 Abstract A general theory of shape which unifies the two classical approaches is organized around two basic intuitions: first, if a boundary were changed only slightly, then in general its shape would change only slightly. This leads to an operational theory of shape based on incremental contour deformations. The second intuition is that not all contours are shapes, but rather only those that can enclose 'physical' material. A novel theory of contour deformation is derived on the basis of abstract conservation principles and Hamilton-Jacobi theory. The result is a characterization of the computational elements of shape: protrusions, parts, bends, and seeds (which show where to place the components of a shape); and leads to a reaction-diffusion space which places shapes within a neighborhood of 'similar' ones. The entropy scale space is obtained from the reaction-diffusion space by running the 'reaction' portion of the equations 'backwards' in time. As a result distinct components of a shape can be removed by introducing a minimal disturbance to the remainder of the shape. As such, the entropy scale space is a combination of smoothing due to shocks as 'black holes' of information and the subsequent rarefaction wave reconstruction and the anisotropic diffusion process spreading of contour information. This technique is numerically stable, and several examples are shown Thesaurus computer vision; convergence of numerical methods; entropy; nonlinear differential equations; partial differential equations; surface topography Other Terms shape approximation; entropy scale space; operational theory of shape; incremental contour deformations; Hamilton-Jacobi theory; computational elements of shape; reaction-diffusion space; rarefaction wave reconstruction; anisotropic diffusion process ClassCodes B6140C; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4495186 AbstractNos. B9311-6140C-153; C9311-5260B-074 ISSN 0277786X References 50 U.S. Copyright Clearance Center Code 0 8194 1280 5/93/$6.00 Country Pub. USA date 1216 ------------------------------------------------------------ Author Andersen, J.D.; Dept. of Comput. Sci., Copenhagen Univ., Denmark Title Methods for modeling the first layers of the retina Source RNNS/IEEE Symposium on Neuroinformatics and Neurocomputers (Cat. NO.92TH0483-8); Part: Rostov-on-Don, Russia; Part: 7-10 Oct. 1992; Sponsored by: IEEE; Russian Neural Networks Soc; new York, NY, USA; IEEE; 2 vol. xxi+1270; 1992; pp. 179-86 vol.1 Abstract Recently, analog preprocessing circuits modeling the first layers of the primate retina (low level vision) have been described. The author compares the operation of these circuits with traditional image preprocessing methods. Specifically, he investigates properties and limitations of a machine vision based edge detection method (inhomogeneous anisotropic diffusion) and makes a comparison with silicon retina architectures. Anisotropic diffusion is a modification of scale space filtering, disallowing smoothing across pronounced image gradients (which are supposed to reflect image edges). The silicon retina is an analog VLSI realization mimicking the function of the first layers of the primate retina. The silicon retina is shown to be faster and to provide better segmentation results than diffusion methods Thesaurus analogue processing circuits; biology computing; computer vision ; digital signal processing chips; edge detection; eye; image segmentation; linear integrated circuits; vision; VLSI; zoology Other Terms analog preprocessing circuits; primate retina; low level vision; image preprocessing methods; machine vision based edge detection method; inhomogeneous anisotropic diffusion; scale space filtering; image gradients; analog VLSI; segmentation; Si retina architectures ClassCodes A8732E; A4230V; B6140C; B1285; B2570; C5260B; C7330; C5160; C5135 Article Type Practical; Experimental Chemical Si/el Language English RecordType Conference ControlNo. 4493808 AbstractNos. A9321-8732E-012; B9311-6140C-092; C9311-5260B-061 ISBN or SBN 0 7803 0809 3 References 15 Country Pub. USA date 1206 ------------------------------------------------------------ Author Jackway, P.T.; Signal Process. Res. Centre, Queensland Univ. of Technol., Brisbane, Qld., Australia Title Multiscale image processing: a review and some recent developments Source Journal of Electrical and Electronics Engineering, Australia; J. Electr. Electron. Eng. Aust. (Australia); vol.13, no.2; June 1993; pp. 88-98 Abstract Many fundamental operations in signal and image processing rely on a smoothing or regularisation procedure. The degree of smoothing or "scale" needs to be carefully selected for optimum results. Instead of concentrating on image processing at a single scale, "multiscale" techniques have recently been introduced to take advantage of the information present at several different scales in the signal. The concept of a "scale-space" provides a way to relate the information obtained over different scales of filter. The Gaussian filter is the only linear filter to possess the required scale-space properties. However, the Gaussian filter has some undesirable properties in two, and higher, dimensions thus limiting the usefulness of Gaussian scale-space. The author presents an overview of some recently developed nonlinear filters which have good scale-space properties in higher dimensions. These developments may open this field up to further study Thesaurus filtering and prediction theory; image processing; mathematical morphology; reviews Other Terms multiscale image processing; review; smoothing; Gaussian filter ; nonlinear filters; scale-space properties; morphological filter; regularisation procedure ClassCodes B6140C; C1250 Article Type Bibliography/Literature Suvery; General or Review Coden JEEADG Language English RecordType Journal ControlNo. 4489122 AbstractNos. B9311-6140C-053; C9311-1250-032 ISSN 07252986 References 63 Country Pub. Australia date 1215 ------------------------------------------------------------ Author Deriche, R.; Giraudon, G.; INRIA Sophia Antipolis, Valbonne, France Title A computational approach for corner and vertex detection Source International Journal of Computer Vision; Int. J. Comput. Vis. (Netherlands); vol.10, no.2; April 1993; pp. 101-24 Abstract Corners and vertexes are strong and useful features in computer vision for scene analysis, stereo matching, and motion analysis. The authors deal with the development of a computational approach to these important features. They consider first a corner model and study analytically its behavior once it has been smoothed using the well-known Gaussian filter. This allows them to clarify the behavior of some well-known cornerness measure based approaches used to detect these points of interest. Most of these classical approaches appear to detect points that do not correspond to the exact position of the corner. A new scale-space based approach that combines useful properties from the Laplacian and Beaudet's measure (1978) is then proposed in order to correct and detect exactly the corner position. An extension of this approach is then developed to solve the problem of trihedral vertex characterisation and detection. In particular, it is shown that a trihedral vertex has two elliptic maxima on extremal contrast surfaces if the contrast is sufficient, and this allows the authors to classify trihedral vertexes in two classes: 'vertex', and 'vertex as corner'. The corner-detection approach developed is applied to accurately detect trihedral vertexes using an additional test in order to make a distinction between trihedral vertexes and corners. Many experiments have been carried out using noisy synthetic data and real images containing corners and vertexes. Most of the promising results obtained are used to illustrate the experimental section Thesaurus computational geometry; computer vision Other Terms corner detection; computational approach; vertex detection; computer vision; scene analysis; stereo matching; motion analysis; Gaussian filter; scale-space based approach; Beaudet's measure; trihedral vertex characterisation; elliptic maxima; extremal contrast surfaces ClassCodes C5260B; C1250; C4260 Article Type Practical Coden IJCVEQ Language English RecordType Journal ControlNo. 4476334 AbstractNos. C9310-5260B-088 ISSN 09205691 References 35 Country Pub. Netherlands date 1213 ------------------------------------------------------------ Author Whitten, G.; Martin Marietta Lab., Balitmore, MD, USA Title Scale space tracking and deformable sheet models for computational vision Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.15, no.7; July 1993; pp. 697-706 Abstract The deformable sheet, a physical model that provides a natural framework for addressing many vision problems that can be solved by smoothness-constrained optimization, is described. Deformable sheets are characterized by a global energy functional, and the smoothness constraint is represented by a linear internal energy term. Analogous to physical sheets, the model sheets are deformed by problem-specific external forces and, in turn, impose smoothness on the applied forces. The model unifies the properties of scale and smoothness into a single parameter that makes it possible to perform scale space tracking by properly controlling the smoothness constraint. Specifically, the desired scale space trajectory is found by solving a differential equation in scale. The simple analytic dependence on scale also provides a mechanism for adaptive step size control. Results from application of the deformable sheet model to various problems in computational vision are presented Thesaurus computer vision; image processing; optimisation Other Terms deformable sheet models; computational vision; smoothness- constrained optimization; global energy functional; smoothness constraint; linear internal energy; scale space tracking; differential equation ClassCodes B6140C; B0260; C1250; C1180 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4465236 AbstractNos. B9310-6140C-015; C9310-1250-013 ISSN 01628828 References 21 U.S. Copyright Clearance Center Code 0162-8828/93/$03.00 Country Pub. USA date 1216 ------------------------------------------------------------ Author Gauch, J.M.; Pizer, S.M.; Coll. of Comput. Sci., Northeastern Univ., Boston, MA, USA Title Multiresolution analysis of ridges and valleys in grey-scale images Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.15, no.6; June 1993; pp. 635-46 Abstract Two methods for identifying and analyzing the multiresolution behavior of ridges and valleys in grey-scale images are described. The first method uses the tools of differential geometry to focus on local image behavior. The resulting vertex curves mark the tops of ridges and bottoms of valleys in an image. The second method focuses on the global drainage patterns of rainfall on a terrain map. The resulting watershed boundaries also identify the tops of ridges and bottoms of valleys in an image. By following these two geometric representations through scale space, the authors build resolution hierarchies on ridges and valleys in the image that can be utilized for interactive image segmentation Thesaurus differential geometry; geophysical techniques; geophysics computing; image segmentation; rain; remote sensing; topography (Earth) Other Terms remote sensing; geophysical techniques; ridges; valleys; grey- scale images; multiresolution behavior; differential geometry; vertex curves; global drainage patterns; rainfall; terrain map; watershed boundaries; interactive image segmentation ClassCodes A9385; A9365; A9110J; A9240E; B7710; B6140C; C7340; C5260B Article Type Practical Coden ITPIDJ Language English RecordType Journal ControlNo. 4465230 AbstractNos. A9319-9385-009; B9310-7710-004; C9310-7340-013 ISSN 01628828 References 28 U.S. Copyright Clearance Center Code 0162-8828/93/$03.00 Country Pub. USA date 1215 ------------------------------------------------------------ Author Crespo, J.; Schafer, R.W.; Sch. of Electr. Eng., Georgia Inst. of Tech., Atlanta, GA, USA Title Image partition using an iterative multi-resolution smoothing algorithm Source ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech and Signal Processing (Cat. No.92CH3103-9); Part: San Francisco, CA, USA; Part: 23-26 March 1992; Sponsored by: IEEE; New York, NY, USA; IEEE; 5 vol. 3219; 1992; pp. 561-4 vol.3 Abstract An iterative algorithm for computing a family of piecewise- constant approximations of images in scale space with variable resolution is presented. The algorithm is a modification of an iterative adaptive smoothing algorithm proposed by P. Saint-Marc et al. (1991). The modification improves the behavior of the algorithm when weak edges are present in the image. Edge information computed from the image is incorporated into the adaptive smoothing approach. The added iterative step does not introduce a significant amount of computation, and the employment of only local operators makes it well suited for a parallel hardware implementation. The edges extracted must be two pixels wide and exhibit high connectivity. A morphological edge operator is presented to fulfil these requirements Thesaurus edge detection; image segmentation; iterative methods; mathematical morphology Other Terms edge extraction; image partition; multi-resolution smoothing algorithm; iterative algorithm; piecewise-constant approximations; scale space; variable resolution; adaptive smoothing; morphological edge operator ClassCodes B6140C; B0290F; C1250; C4130 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4454548 AbstractNos. B9309-6140C-123; C9309-1250-110 ISBN or SBN 0 7803 0532 9 References 4 U.S. Copyright Clearance Center Code 0 7803 0532 9/92/$3.00 Country Pub. USA date 1199 ------------------------------------------------------------ Author Brockett, R.W.; Maragos, P.; Div. of Appl. Sci., Harvard Univ., Cambridge, MA, USA Title Evolution equations for continuous-scale morphology Source ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech and Signal Processing (Cat. No.92CH3103-9); Part: San Francisco, CA, USA; Part: 23-26 March 1992; Sponsored by: IEEE; New York, NY, USA; IEEE; 5 vol. 3219; 1992; pp. 125-8 vol.3 Abstract Several nonlinear partial differential equations that model the scale evolution associated with continuous-space multiscale morphological erosions, dilations, openings, and closings are discussed. These systems relate the infinitesimal evolution of the multiscale signal ensemble in scale space to a nonlinear operator acting on the space of signals. The type of this nonlinear operator is determined by the shape and dimensionality of the structuring element used by the morphological operators, generally taking the form of nonlinear algebraic functions of certain differential operators Thesaurus image processing; mathematical morphology; nonlinear differential equations Other Terms evolution equations; continuous-scale morphology; nonlinear partial differential equations; scale evolution; multiscale morphological erosions; dilations; openings; closings; multiscale signal ensemble; scale space; nonlinear operator; structuring element; morphological operators; nonlinear algebraic functions; differential operators ClassCodes B6140C; B0220 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4445906 AbstractNos. B9308-6140C-203 ISBN or SBN 0 7803 0532 9 References 11 U.S. Copyright Clearance Center Code 0 7803 0532 9/92/$3.00 Country Pub. USA date 1199 ------------------------------------------------------------ Author Acton, S.T.; Bovik, A.C.; Dept. of Electr. & Comput. Eng., Texas Univ., Austin, TX, USA Title Anisotropic edge detection using mean field annealing Source ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech and Signal Processing (Cat. No.92CH3103-9); Part: San Francisco, CA, USA; Part: 23-26 March 1992; Sponsored by: IEEE; New York, NY, USA; IEEE; 5 vol. 3219; 1992; pp. 393-6 vol.2 Abstract An edge detection technique that optimizes edge localization while providing edge continuity and edge thinning is introduced. The solution is obtained by annealing a mean field neural network, providing inexpensive solutions with high parameter insensitivity. Anisotropic diffusion is used to provide localized edge data through the scale-space. Analysis of network parameters, diffusion parameters, network convergence, and scale-space equivalence is provided. Results are shown for real image data and compared with the results of other important edge detection schemes Thesaurus edge detection; neural nets Other Terms anisotropic edge detection; anisotropic diffusion; mean field annealing; edge localization; edge continuity; edge thinning; mean field neural network; parameter insensitivity; localized edge data; scale-space; network parameters; diffusion parameters; network convergence; real image data ClassCodes B6140C; C1250; C1230D Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4441475 AbstractNos. B9308-6140C-145; C9308-1250-120 ISBN or SBN 0 7803 0532 9 References 6 U.S. Copyright Clearance Center Code 0 7803 0532 9/92/$3.00 Country Pub. USA date 1199 ------------------------------------------------------------ Author Wong, Y.; Posner, E.C.; Dept. of Electr. Eng., California Inst. of Technol., Pasadena, CA, USA Title Scale-space clustering and classification of SAR images with numerous attributes and classes Source Proceedings. IEEE Workshop on Applications of Computer Vision (Cat. No.92TH0446-5); Part: Palm Springs, CA, USA; Part: 30 Nov. -2 Dec. 1992; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xi+317; 1992; pp. 74-81 Abstract Describes application of scale-space clustering to the classification of a multispectral and polarimetric SAR image of an agricultural site. After polarimetric and radiometric calibration and noise cancellation, the authors extracted a 12- dimensional feature vector for each pixel from the scattering matrix. The algorithm was able to partition without supervision a set of unlabeled vectors from 13 selected sites, each site corresponding to a distinct crop, into 13 clusters. The cluster parameters were then used to classify the whole image. The classification map is much less noisy and more accurate than those obtained by hierarchical rules. The algorithm can handle variabilities in cluster densities, cluster sizes and ellipsoidal shapes Thesaurus agriculture; image recognition; remote sensing by radar; synthetic aperture radar Other Terms scale-space clustering; SAR image; agricultural site; classification map; variabilities; cluster densities ClassCodes B7730; B6300; C5260B Article Type Practical Language English RecordType Conference ControlNo. 4437615 AbstractNos. B9308-7730-003; C9308-5260B-062 ISBN or SBN 0 8186 2840 5 References 20 U.S. Copyright Clearance Center Code 0 8186 2840 5/92/$03.00 Country Pub. USA date 1207 ------------------------------------------------------------ Author Rangarajan, K.; Allen, W.; Shah, M.; Dept. of Comput. Sci., Univ. of Central Florida, Orlando, FL, USA Title Matching motion trajectories using scale-space Source Pattern Recognition; Pattern Recognit. (UK); vol.26, no.4; April 1993; pp. 595-610 Abstract The goal is to design a recognition system which can distinguish between two objects with the same shape but different motion, or between two objects with the same motion but a different shape. The input to the system is a set of two-dimensional (2D) trajectories from an object tracked through a sequence of n frames. The structure and three-dimensional (3D) trajectories of each object in the domain are stored in the model. The problem is to match the information in the model with the input set of 2D trajectories and determine if they represent the same object. The simplest way to perform these steps is to match the input 2D trajectories with the 2D projections of the 3D model trajectories. First, a simple algorithm is presented which matches two single trajectories using only motion information. The 2D motion trajectories are converted into two one-dimensional (1D) signals based on their speed and direction components. The signals are then represented by scale-space images, both to simplify matching and because the scale-space representations are translation and rotation invariant. The matching algorithm is extended to include spatial information and a second algorithm is proposed which matches multiple trajectories by combining motion and spatial match scores. Both algorithms are tested with real and synthetic data Thesaurus image recognition; motion estimation Other Terms 2D trajectories; translation invariant; 3D trajectories; speed components; motion trajectories; scale-space; recognition system; shape; input set; model trajectories; direction components; rotation invariant; spatial information ClassCodes C5260B; C1250 Article Type Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 4436481 AbstractNos. C9308-5260B-036 ISSN 00313203 References 8 U.S. Copyright Clearance Center Code 0031-3203/93/$6.00+.00 Country Pub. UK date 1213 ------------------------------------------------------------ Author Mori, K.; Doi, S.; Matsuda, M.; Osaka Electro-Commun. Univ., Japan Title The approach of the real-time underwater data transmission system based on scale-space filtering Source Transputer/Occam. Japan 4. Proceedings of the 4th Transputer/Occam International Conference; Part: Tokyo, Japan; Pa rt: 4-5 June 1992; Sponsored by: Inmos; Nikkan Kogyo Shimbun; Amsterdam, Netherlands; IOS Press; ix+268; 1992; pp. 121-36 Editor Noguchi, S.; Umeo, H. Abstract Underwater data transmission using scale-space filtering and fingerprint pattern keying as the data transmission method is proposed. The design of the real-time underwater data transmission system is presented. The matching method using the number of the zero-cross points in fingerprint pattern keying is also discussed as one of the data transmission methods which is conscious of real-time processing. From two types of simulations on a transputer array, where one is the division by the scale parameter sigma /sub j/ and the other is pipeline-processing using only one scale parameter, the possibility about the real- time processing of this system is discussed using experimental results which are the execution time on a transputer array Thesaurus acoustic signal processing; codes; data communication systems; filtering and prediction theory; parallel algorithms; real-time systems; transputer systems Other Terms real-time underwater data transmission system; scale-space filtering; fingerprint pattern keying; matching method; zero- cross points; transputer array; scale parameter; pipeline- processing ClassCodes B6140; B6120B; B6210; C5620; C5260; C1260; C6130 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 4412649 AbstractNos. B9307-6140-046; C9307-5620-049 References 7 Country Pub. Netherlands date 1202 ------------------------------------------------------------ Author Blom, J.; Ter Haar Romeny, B.M.; Bel, A.; Koenderink, J.J.; Buys Ballot Lab., Utrecht Univ., Netherlands Title Spatial derivatives and the propagation of noise in Gaussian scale space Source Journal of Visual Communication and Image Representation; J. Vis. Commun. Image Represent. (USA); vol.4, no.1; March 1993; pp. 1-13 Abstract Image structure analysis requires the computation of local spatial derivatives of the intensity distribution. These are determined by convolution with Gaussian derivative operators. Noise is enhanced by the high-pass nature of differentiation, particularly at high order. On the other hand, the Gaussian- weighted averaging gives rise to noise reduction. The author gives an analysis of the propagation of spatially uncorrelated as well as spatially correlated additive noise in scale space, when the noise is subjected to fuzzy derivative operators of any order. The propagation of noise variance is always substantially reduced when scale is increased, the effect being greater for higher order derivatives. The spatial blurring is always predominant, or, the representation of the noise and its derivatives is substantial only at the original (inner) scale. Expressions are derived for the propagation of noise in functions of derivatives, like the Laplacian and isophote curvature. All expressions are evaluated for a D-dimensional (image) data structure. Determining derivatives, even up to high order, combined with scale space, is a very robust and stable operation. The important conclusion is that the use of differential geometrical methods in scale space, particularly in noisy images, is justified Thesaurus correlation theory; edge detection; filtering and prediction theory; image processing; noise; spatial filters Other Terms noise propagation; image structure analysis; Laplacian operator; edge detection; filtering; Gaussian scale space; local spatial derivatives; intensity distribution; spatially correlated additive noise; fuzzy derivative operators; spatial blurring; isophote curvature ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden JVCRE7 Language English RecordType Journal ControlNo. 4406258 AbstractNos. B9306-6140C-231; C9306-1250-189 ISSN 10473203 References 22 U.S. Copyright Clearance Center Code 1047-3203/93/$5.00 Country Pub. USA date 1212 ------------------------------------------------------------ Author Kube, P.; Dept. of Comput. Sci. & Eng., California Univ., San Diego, La Jolla, CA, USA Title Properties of energy edge detectors Source Proceedings. 1992 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.92CH3168-2); Part: Champaign, IL, USA; Part: 15-18 June 1992; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xvi+870; 1992; pp. 586-91 Abstract The author introduces a framework for investigating the properties of energy edge detectors and uses it to derive some results of interest. He shows a necessary condition on the form of constituent linear filters in quadratic detectors, subject to some conditions, and demonstrates some limitations of such detectors. It is shown that no quadratic detector can detect an edge at 0 for both a sinewave and a cosine wave, which has implications for detecting narrowband edges with spatially local filters. It is also shown that the scale-space behavior of energy detectors is not well-behaved, in that it contains bifurcations as scale increases, i.e. new edges can be created as the image is smoothed Thesaurus image segmentation; pattern recognition Other Terms energy edge detectors properties; necessary condition; constituent linear filters; quadratic detectors; sinewave; cosine wave; spatially local filters; scale-space behavior; bifurcations ClassCodes B6140C; C1250; C5260B Article Type Practical Language English RecordType Conference ControlNo. 4399892 AbstractNos. B9306-6140C-189; C9306-1250-150 ISBN or SBN 0 8186 2855 3 References 12 U.S. Copyright Clearance Center Code 0 8186 2855 3/92$03.00 Country Pub. USA date 1202 ------------------------------------------------------------ Author Whitaker, R.T.; Pizer, S.M.; Dept. of Comput. Sci., North Carolina Univ., Chapel Hill, NC, USA Title A multi-scale approach to nonuniform diffusion Source CVGIP: Image Understanding; CVGIP, Image Underst. (USA); vol.57, no.1; Jan. 1993; pp. 99-110 Abstract The paper examines a new method of image processing that combines information at multiple scales in order to locate boundaries. The method employs a technique of edge-affected diffusion, where blurring is limited by the presence of edges as measured at the scale of interest. By repeating such processing and measuring gradients at successively smaller one is able to trace a 'path' through scale space which can preserve accurate information about boundaries of objects, and yet selectively remove objects that fall below a scale of interest. The authors compared this approach with the anisotropic diffusion technique described by Perona and Malik (IEEE Trans. Pattern Anal. Mach. Intell. vol.12, p.429-39, 1990), which depends only on the local gradient of intensity of the processed image. The authors show some examples which indicate that this method could be useful for boundary detection in the presence of blurring and noise and which is also capable of performing grouping of distinct objects at various scales. The paper also examines the sensitivity of this process with respect to one's choice of parameters Thesaurus edge detection Other Terms image processing; edge-affected diffusion; blurring; boundary detection ClassCodes B6140C; C5260B; C1250; C1260 Article Type Theoretical / Mathematical; Experimental Coden CIUNEJ Language English RecordType Journal ControlNo. 4398172 AbstractNos. B9306-6140C-150; C9306-5260B-122 ISSN 10499660 References 13 U.S. Copyright Clearance Center Code 1049-9660/93/$5.00 Country Pub. USA date 1210 ------------------------------------------------------------ Author Raman, S.V.; Sarkar, S.; Boyer, K.L.; Dept. of Electr. Eng., Ohio State Univ., Columbus, OH, USA Title Hypothesizing structures in edge-focused cerebral magnetic resonance images using graph-theoretic cycle enumeration Source CVGIP: Image Understanding; CVGIP, Image Underst. (USA); vol.57, no.1; Jan. 1993; pp. 81-98 Abstract The authors present a novel method for the automatic generation of structure hypotheses (that is, educated guesses) suitable for recognition in medical images. They base their approach on segment-based edge-focusing to delineate significant boundaries precisely, and graph-theoretic cycle enumeration to produce natural closures and, therefore, plausible tissue structures of interest from incomplete boundary information. An efficient edge focusing algorithm selects significant fine scale boundaries as those natural descendants (in scale space) of prominent coarse scale edges. The fine scale representation provides the localization precision necessary, while the focusing ensures that only significant contours surviving over a range of scales are considered and so eliminates much of the 'clutter' associated with a fine scale edge map. The spatial relationships among the edge segments are stored in the form of a directed graph. Possible extensions (closures) of broken edge segments are searched using time- and space-efficient voting methods. Cycle enumeration techniques for directed graphs then generate the structure hypotheses Thesaurus biomedical NMR; brain; directed graphs; edge detection; feature extraction; image recognition Other Terms hypothesis generation; edge-focused cerebral magnetic resonance images; graph-theoretic cycle enumeration; medical images; segment-based edge-focusing; tissue structures; incomplete boundary information; edge focusing algorithm; fine scale boundaries; coarse scale edges; fine scale representation; localization precision; significant contours; spatial relationships; edge segments; directed graph; broken edge segments; voting methods; structure hypotheses ClassCodes A8770E; A8740; A8760G; B6140C; B7510B; C7330; C5260B; C1160 Article Type Practical; Theoretical / Mathematical; Experimental Coden CIUNEJ Language English RecordType Journal ControlNo. 4398171 AbstractNos. A9311-8770E-015; B9306-6140C-149; C9306-7330-049 ISSN 10499660 References 36 U.S. Copyright Clearance Center Code 1049-9660/93/$5.00 Country Pub. USA date 1210 ------------------------------------------------------------ Author Xu Zhi Xiang; Wang Jijie; Shangai Univ. of Sci. & Technol., China Title Performances of the Laplacian of binomial distribution and the discrete Laplacian of Gaussian edge detection operators Source Acta Electronica Sinica; Acta Electron. Sin. (China); vol.20, no.11; Nov. 1992; pp. 69-74 Abstract Presents a performance analysis and comparison between the Laplacian of binomial distribution (LOB) and the discrete Laplacian of Gaussian (DLOG) edge detection operators in the space domain and frequency domain. When the scale space constants of the two edge detection operators are large enough, the characteristics in space domain and frequency domain and performances in image edge detection are almost the same. But when the scale space constants are smaller, the conclusion can be made that the performances of LOB operator are little better than that of discrete LOG operator after comparisons of central frequency, 3dB bandwidth and high frequency attenuation rate at cut off frequency in frequency domain. LOB operator may be considered as a discrete realization of LOG operator. The experiments are given to verify the correctness of analysis Thesaurus edge detection Other Terms Laplacian; binomial distribution; discrete Laplacian; Gaussian edge detection operators; performance analysis; space domain; frequency domain; scale space constants; image edge detection; LOB operator; discrete LOG operator; central frequency; 3dB bandwidth; frequency attenuation rate; cut off frequency; correctness ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden TTHPAG Language Chinese RecordType Journal ControlNo. 4396136 AbstractNos. B9306-6140C-137; C9306-1250-106 ISSN 03722112 References 3 Country Pub. China date 1207 ------------------------------------------------------------ Author Wong, Y.-F.; Dept. of Electr. Eng., California Inst. of Technol., Pasadena, CA, USA Title Clustering data by melting Source Neural Computation; Neural Comput. (USA); vol.5, no.1; Jan. 1993; pp. 89-104 Abstract The author derives a new clustering algorithm based on information theory and statistical mechanics, which is the only algorithm that incorporates scale. It also introduces a new concept into clustering: cluster independence. The cluster centers correspond to the local minima of a thermodynamic free energy, which are identified as the fixed points of a one- parameter nonlinear map. The algorithm works by melting the system to produce a tree of clusters in the scale space. Melting is also insensitive to variability in cluster densities, cluster sizes, and ellipsoidal shapes and orientations. The authors tested the algorithm successfully on both simulated data and a synthetic aperture radar image of an agricultural site with 12 attributes for crop identification Thesaurus information theory; pattern recognition; remote sensing by radar ; statistical mechanics Other Terms clustering algorithm; information theory; statistical mechanics; cluster independence; cluster centers; local minima; thermodynamic free energy; one-parameter nonlinear map; melting; scale space; cluster densities; cluster sizes; ellipsoidal shapes; orientations; synthetic aperture radar image; agricultural site; crop identification ClassCodes C1260; C1250; C1180 Article Type Theoretical / Mathematical Coden NEUCEB Language English RecordType Journal ControlNo. 4393900 AbstractNos. C9306-1260-022 ISSN 08997667 References 24 Country Pub. USA date 1210 ------------------------------------------------------------ Author Eggert, D.W.; Bowyer, K.W.; Dyer, C.R.; Christensen, H.I.; Goldgof, D.B.; Dept. of Comput. Sci. & Eng., Univ. of South Florida, Tampa, FL, USA Title The scale space aspect graph Source Proceedings. 1992 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.92CH3168-2); Part: Champaign, IL, USA; Part: 15-18 June 1992; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xvi+870; 1992; pp. 335-40 Abstract Currently the aspect graph is computed under the assumption of perfect resolution in the viewpoint, the projected image, and the object shape. Visual detail is represented that an observer might never see in practice. By introducing scale into this framework, a mechanism is provided for selecting levels of detail that are large enough to merit explicit representation, effectively allowing control over the size of the aspect graph. To this end the scale space aspect graph is introduced, and an interpretation of the scale dimension in terms of the spatial extent of image features is considered. A brief example is given for polygons in a plane Thesaurus graphs; image recognition Other Terms detail representation; scale space aspect graph; perfect resolution; projected image; object shape; levels of detail; image features; polygons ClassCodes C5260B; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4382947 AbstractNos. C9305-5260B-085 ISBN or SBN 0 8186 2855 3 References 22 U.S. Copyright Clearance Center Code 0 8186 2855 3/92$03.00 Country Pub. USA date 1202 ------------------------------------------------------------ Author Fermuller, C.; Kropatsch, W.; Comput. Vision Lab., Maryland Univ., College Park, MD, USA Title Multi-resolution shape description by corners Source Proceedings. 1992 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.92CH3168-2); Part: Champaign, IL, USA; Part: 15-18 June 1992; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xvi+870; 1992; pp. 271-6 Abstract A robust method for describing planar curves in multiple resolution using curvature information is presented. The method is developed by taking into account the discrete nature of digital images as well as the discrete aspect of a multiresolution structure (pyramid). The main contribution lies in the robustness of the technique, which is due to the additional information that is extracted from observing the behavior of corners in the whole pyramid. Furthermore, the resulting algorithm is conceptually simple and easily parallelizable. Theoretical results are developed analyzing the curvature of continuous curves in scale-space and showing the behavior of curvature extrema under varying scale. The results are used to eliminate any ambiguities that might arise from sampling problems due to the discreteness of the representation. Experimental results demonstrate the potential of the method Thesaurus computational geometry; computer vision; curve fitting; feature extraction; image processing Other Terms planar curves; multiple resolution; curvature information; digital images; multiresolution structure; robustness; corners; parallelizable; continuous curves; scale-space; curvature extrema; varying scale; ambiguities ClassCodes C5260B; C4260 Article Type Experimental Language English RecordType Conference ControlNo. 4382937 AbstractNos. C9305-5260B-075 ISBN or SBN 0 8186 2855 3 References 16 U.S. Copyright Clearance Center Code 0 8186 2855 3/92$03.00 Country Pub. USA date 1202 ------------------------------------------------------------ ------------------------------------------------------------ Author Mjolsness, E.; Dept. of Comput. Sci., Yale Univ., New Haven, CT, USA Title Visual grammars and their neural networks Source Science of Artificial Neural Networks; Part: Orlando, FL, USA; Pa rt: 21-24 April 1992; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1710, pt.1; 1992; pp. 63-85 vol.1 Abstract Exhibits a systematic way to derive neural nets for vision problems. It involves formulating a vision problem as Bayesian inference or decision on a comprehensive model of the visual domain given by the probabilistic grammar. A key feature of this grammar is the way in which it eliminates model information, such as object labels, as it produces an image; correspondance problems and other noise removal tasks result. The neural nets that arise most directly are generalized assignment networks. Also there are transformations which naturally yield improved algorithms such as correlation matching in scale space and the Frameville neural nets for high-level vision. Networks derived this way generally have objective functions with spurious local minima; such minima may commonly be avoided by dynamics that include deterministic annealing, for example recent improvements to Mean Field Theory dynamics. The grammatical method of neural net design allows domain knowledge to enter from all levels of the grammar, including 'abstract' levels remote from the final image data, and may permit new kinds of learning as well Thesaurus computer vision; grammars; inference mechanisms; neural nets Other Terms neural nets; vision problems; Bayesian inference; probabilistic grammar; domain knowledge ClassCodes C1230D; C1250; C4210; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4374026 AbstractNos. C9305-1230D-044 ISSN 0277786X References 27 U.S. Copyright Clearance Center Code 0 8194 0875 1/92/$4.00 Country Pub. USA date 1200 ------------------------------------------------------------ Author Trucco, E.; Dept. of Artificial Intelligence, Edinburgh Univ., UK Title On shape-preserving boundary conditions for diffusion smoothing Source Proceedings. 1992 IEEE International Conference on Robotics And Automation (Cat. No.92CH3140-1); Part: Nice, France; Part: 12- 14 May 1992; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 3 vol. xxxix+2819; 1992; pp. 1690-4 vol.2 Abstract Several boundary treatments for attenuating shape deformation introduced by Gaussian smoothing are discussed. The author models Gaussian smoothing by a diffusion equation, a general mathematical framework particularly useful for scale-space analysis. An adaptive, shape-reserving boundary condition for diffusion smoothing range images is introduced. It is shown that this condition is more general than similar techniques found in the literature Thesaurus adaptive systems; filtering and prediction theory; image processing; sensor fusion Other Terms image analysis; shape-preserving boundary conditions; diffusion smoothing; shape deformation; Gaussian smoothing; scale-space analysis; range images ClassCodes B6140C; C1250; C1260; C1240 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4368196 AbstractNos. B9304-6140C-182; C9304-1250-137 ISBN or SBN 0 8186 2720 4 References 15 U.S. Copyright Clearance Center Code 0 8186 2720 4/92/$03.00 Country Pub. USA date 1201 ------------------------------------------------------------ ------------------------------------------------------------ Author Woo Young Choi; Rae-Hong Park; Dept. of Electr. Eng., Myongji Univ., Seoul, South Korea Title Stereo matching using dynamic programming in scale-space Source Journal of the Korean Institute of Telematics and Electronics; J. Korean Inst. Telemat. Electron. (South Korea); vol.29B, no.8; Aug. 1992; pp. 44-53 Abstract In this paper, a matching method is proposed to improve the correct matching rate in stereo correspondence matching in which the fingerprint of zero-crossing points on the scale-space is used as the robust matching feature. Dynamic programming, which is appropriate for the fingerprint feature, is introduced for correspondence matching. The authors also improve the matching rate by using the post-processing for correcting mismatched points. In simulation, they apply the proposed algorithm to the synthetic and real images and obtain good matching results Thesaurus dynamic programming; feature extraction; image recognition Other Terms synthetic images; dynamic programming; scale-space; matching rate; correspondence matching; fingerprint; zero-crossing points; robust matching feature; post-processing; real images ClassCodes B6140C; B0260; C1250; C1180 Article Type Theoretical / Mathematical Coden CKNOEZ Language Korean RecordType Journal ControlNo. 4365921 AbstractNos. B9304-6140C-154; C9304-1250-118 ISSN 1016135X References 12 Country Pub. South Korea date 1204 ------------------------------------------------------------ Author Soo-Chang Pei; Chao-Nan Lin; Dept. of Electr. Eng., Nat. Taiwan Univ., Taipei, Taiwan Title The detection of dominant points on digital curves by scale-space filtering Source Pattern Recognition; Pattern Recognit. (UK); vol.25, no.11; Nov. 1992; pp. 1307-14 Abstract An effective method of scale-space filtering with a Gaussian kernel is introduced to detect dominant points on digital curves. The conventional polygonal approximation algorithms are time- consuming and need input parameter tuning for Gaussian smoothing the noise and quantization error, also they are sensitive to scaling and rotation of the object curve. The above difficulty can be overcome by finding out the dominant points at each scale by scale-space filtering. By tracing back the dominant point contours in the scale-space image, the stable cardinal curvature points can be detected very accurately. This new method requires no input parameters, and the resultant dominant points do not change under translation, rotation and scaling. A fast convolution algorithm is proposed to detect the dominant points at each scale Thesaurus filtering and prediction theory; image recognition Other Terms dominant point detection; Gaussian smoothing; image recognition; digital curves; scale-space filtering; scale-space image; stable cardinal curvature points; fast convolution algorithm ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 4361821 AbstractNos. B9304-6140C-124; C9304-1250-095 ISSN 00313203 References 19 U.S. Copyright Clearance Center Code 0031-3203/92/$5.00+.00 Country Pub. UK date 1207 ------------------------------------------------------------ Author de Ridder, H.; Inst. for Perception Res., Eindhoven, Netherlands Title Minkowski-metrics as a combination rule for digital-image-coding impairments Source Human Vision, Visual Processing and Digital Display III; Part: San Jose, CA, USA; Part: 10-13 Feb. 1992; Sponsored by: SPIE; Soc. Imaging Sci. Technol; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1666; 1992; pp. 16-26 Abstract The urge to compress the amount of information needed to represent digitized images while preserving perceptual image quality has led to a plethora of image-coding algorithms. At high data compression ratios, these algorithms usually introduce several coding artifacts. For the evaluation of image-coding algorithms, it is important to find out how these impairments combine and how this can be described. The objective is to show that Minkowski-metrics can be used as a combination rule for small impairments like those usually encountered in digitally coded images. An experiment has been conducted in which subjects assessed the perceptual quality of scale-space-coded color images comprising three kinds of impairment, namely 'unsharpness', 'phantoms' (dark/bright patches within bright/dark homogeneous regions) and 'color desaturation'. The results show an accumulation of these impairments that is efficiently described by a Minkowski-metric with an exponent of about two Thesaurus colour vision; image coding; visual perception Other Terms digital-image-coding impairments; perceptual image quality; plethora; data compression; Minkowski-metrics; scale-space- coded color images; unsharpness; phantoms; color desaturation ClassCodes A4230V; A8732S; A8732N; B6140C; B6120B; C1250; C5260B Article Type Experimental Coden PSISDG Language English RecordType Conference ControlNo. 4337828 AbstractNos. A9306-4230-002; B9303-6140C-137; C9303-1250-126 ISSN 0277786X References 19 U.S. Copyright Clearance Center Code 0 8194 0820 4/92/$4.00 Country Pub. USA date 1198 ------------------------------------------------------------ ------------------------------------------------------------ ------------------------------------------------------------ Author Vandenberg, S.; Osborne, C.F.; Dept. of Phys., Monash Univ., Caulfield East, Vic., Australia Title Digital image processing techniques, fractal dimensionality and scale-space applied to surface roughness Source Wear; Wear (Switzerland); vol.159, no.1; 2 Nov. 1992; pp. 17-30 Abstract Surface roughness is a result of the existence of peaks and troughs in the surface. Measurement of the surface topography is controlled by the resolution of the probe (e.g. a diamond stylus, a laser beam) and the literature shows that the predicted topography has great sensitivity to the probe dimensions. Image analysis techniques are applied to the topography of the surface and the properties of the surface are predicted on the assumption that the surface is modellable by two-dimensional random noise. A fractal model is applied to the model of the noise and the limitations of this model are investigated with respect to sampling and digitization. Finally the ideas of 'scale-space' are applied to the images to establish the nature of the basic roughness elements which describe the surfaces under consideration Thesaurus fractals; physics computing; surface topography Other Terms digital image processing technique; fractal dimensionality; surface roughness; surface topography; two-dimensional random noise ClassCodes A6820; C7320 Article Type Theoretical / Mathematical Coden WEARCJ Language English RecordType Journal ControlNo. 4335470 AbstractNos. A9305-6820-022; C9303-7320-089 ISSN 00431648 References 33 U.S. Copyright Clearance Center Code 0043-1648/92/$5.00 Country Pub. Switzerland date 1207 ------------------------------------------------------------ Author Chung-Lin Huang; Tai-Yuen Cheng; Chaur-Chin Chen; Dept. of Electr. Eng., Nat. Tsing-Hua Univ., Hsin-Chu, Taiwan Title Color images' segmentation using scale space filter and Markov random field Source Pattern Recognition; Pattern Recognit. (UK); vol.25, no.10; Oct. 1992; pp. 1217-29 Abstract A hybrid method is presented that combines the scale space filter (SSF) and Markov random field (MRF) for color image segmentation. The fundamental idea of the SSF is to use the convolution of Gaussian functions and an image-histogram to generate a scale space image and then find the proper interval bounded by the local extrema of the derivatives. The Gaussian function is with zero mean and varied standard deviation. Using the SSF the different scaled histogram is separated into intervals corresponding to peaks and valleys. The MRF makes use of the property that each pixel in an image has some relationship with other pixels. The basic construction of an MRF is a joint probability given the original data. The original data is the image that is obtained from the source and the result is called the label image. Because the MRF needs a number of segments before it converges to the global minimum, the SSF is exploited to do coarse segmentation and then MRF is used to do fine segmentation of the images. Basically, the former is histogram- based segmentation, whereas the latter is neighborhood-based segmentation Thesaurus image segmentation; Markov processes; simulated annealing; spatial filters Other Terms spatial filters; simulated annealing; Gibbs sampling; scale space filter; Markov random field; color image segmentation; convolution; Gaussian functions; image-histogram; coarse segmentation; fine segmentation; neighborhood-based segmentation ClassCodes B6140C; B0240Z; B0260; C1250; C1140Z; C1180 Article Type Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 4329058 AbstractNos. B9303-6140C-043; C9303-1250-037 ISSN 00313203 References 18 U.S. Copyright Clearance Center Code 0031-3203/92/$5.00+.00 Country Pub. UK date 1206 ------------------------------------------------------------ Author Goshtasby, A.; Dept. of Electr. Eng. & Comput. Sci., Coll. of Eng., Chicago, IL, USA Title Parametric representation of digital shapes by Gaussian functions Source Computer Aided Design; Comput. Aided Des. (UK); vol.24, no.12; Dec. 1992; pp. 659-65 Abstract A new formulation for parametric curves is given. In the proposed formulation, each component of a curve is defined independently of the other components and with different knots. A curve is represented by a blending of Gaussian functions. The Gaussian functions are estimated by the scale-space analysis of a digital shape. The estimated Gaussian functions are then refined by the Marquardt algorithm to minimize the root-mean-squared error between the curve and the shape. Numerical examples are given showing the accuracy and compression rate of the proposed parametric curve in the representation of digital shapes Thesaurus computational geometry; curve fitting; data compression; image processing Other Terms parametric representation; digital shapes; Gaussian functions; parametric curves; scale-space analysis; digital shape; Marquardt algorithm; root-mean-squared error; compression rate ClassCodes C4260; C4130; C5260B Article Type Practical; Theoretical / Mathematical Coden CAIDA5 Language English RecordType Journal ControlNo. 4326506 AbstractNos. C9303-4260-003 ISSN 00104485 References 30 U.S. Copyright Clearance Center Code 0010-4485/92/120659-07$3.00 Country Pub. UK date 1208 ------------------------------------------------------------ Author Nguyen, T.C.; Huang, T.S.; Beckman Inst. & Coordinated Sci. Lab., Urbana, IL, USA Title Image blurring effects due to depth discontinuities: blurring that creates emergent image details Source Image and Vision Computing; Image Vis. Comput. (UK); vol.10, no.10; Dec. 1992; pp. 689-98 Abstract A new model (called multi-component blurring-or MCB) to account for image blurring effects due to depth discontinuities is presented. It is shown that blurring processes operating in the vicinity of large depth discontinuities can give rise to emergent image details. In other words, the maximum principle for scale space does not hold. It is argued that blurring in high-relief 3D scenes should be more accurately modelled as a multi-component process. Results are presented from extensive and carefully designed experiments, with many images of real scenes taken by a CCD camera with typical parameters. These results have consistently supported the new blurring model. Due care was taken to ensure that the image phenomena observed are mainly due to defocussing and not due to mutual illuminations, specularity, objects' 'finer' structures, coherent diffraction, or incidental image noises. The paper also hypothesizes on the role of blurring on human depth-from-blur perception, based on correlation with recent results from human blur perception Thesaurus computer vision; image recognition Other Terms depth discontinuities; emergent image details; multi-component blurring; image blurring; high-relief 3D scenes; real scenes; CCD camera; defocussing; human depth-from-blur perception ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical Coden IVCODK Language English RecordType Journal ControlNo. 4320317 AbstractNos. B9302-6140C-221; C9302-1250-195 ISSN 02628856 References 20 U.S. Copyright Clearance Center Code 0262-8856/92/010689-10$3.00 Country Pub. UK date 1208 ------------------------------------------------------------ Author Jang, B.K.; Chin, R.T.; Health Sci. Res. Lab., Eastman Kodak Co., Rochester, NY, USA Title Gaussian and morphological scale space for shape analysis Source Asia-Pacific Engineering Journal, Part A (Electrical Engineering); Asia-Pac. Eng. J. A, Electr. Eng. (Singapore); vol.2, no.2; June 1992; pp. 165-202 Abstract Multiscale image representations, or scale space, have been utilized in coarse-to-fine image processing, in which the image is represented by sets of features, each set presented at a different scale. A one-dimensional signal can be represented as a two-dimensional scale space in which feature locations are encoded spatially and their evolution through scale is encoded in the second dimension resulting in two-dimensional line patterns. A two-dimensional image in turn produces a three-dimensional scale space of surfaces. The idea behind scale space is based on the fact that single-scale representations and single-scale processing are inadequate in many applications because an image cannot categorically be assumed to have only features of a single size. Scale space has been successfully applied to applications such as noise filtering, corner detection, and recognition. The construction of scale space requires the smoothing of the given image to generate a set of corresponding images at other coarser scales, and the extraction of features at these scales. Various methods of smoothing combined with various feature extractors will result in drastically different scale space representations. The particular application and desired criteria determine the choice of smoothing and feature extractor. This paper reviews the Gaussian and morphological scale space for planar shape analysis. The discussion of the various scale space methods is organized into three categories-boundary approach, region approach and hybrid approach. Properties, limitations, performance and applications of these scale space methods are discussed. A number of examples are given to illustrate the various essential properties and problems associated with scale space Thesaurus feature extraction; filtering and prediction theory; image recognition Other Terms morphological scale space; shape analysis; coarse-to-fine image processing; one-dimensional signal; two-dimensional scale space; feature locations; two-dimensional line patterns; three- dimensional scale space; noise filtering; corner detection; smoothing; boundary approach; region approach; hybrid approach ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden APEJEM Language English RecordType Journal ControlNo. 4317702 AbstractNos. B9302-6140C-204; C9302-1250-184 ISSN 01295411 References 277 Country Pub. Singapore date 1202 ------------------------------------------------------------ Author Morita, S.; Kawashima, T.; Aoki, Y.; Fac. of Eng., Hokkaido Univ., Sapporo, Japan Title Generation of hierarchical description for smooth curved surfaces Source Transactions of the Institute of Electronics, Information and Communication Engineers D-II; Trans. Inst. Electron. Inf. Commun. Eng. D-II (Japan); vol.J75D-II, no.8; Aug. 1992; pp. 1353-63 Abstract Introduces a method to create a hierarchical description of smooth curved surfaces with scale-space analysis. The authors extend the scale-space method used in 1D signal analysis to a 3D surface. An object is described with two steps. In the first step, a 3D scale-space images are segmented by zero-crossings of curvatures at each scale and then linked between consecutive scales based on topological changes (KH-description). In the second step, the KH-description is parsed and translated into a PS-tree which contains the number and distribution of regions required for shape matching. The multiresolution hierarchical description contains coarse-to-fine shape information of the object, and examples show that the PS-tree symbolic description is suitable for efficient coarse-to-fine 3D shape matching. The effectiveness of this approach is demonstrated for raw data derived from a range finder Thesaurus feature extraction; image segmentation; topology; trees (mathematics) Other Terms 3D image segmentation; smooth curved surfaces; scale-space analysis; zero-crossings; curvatures; topological changes; KH- description; PS-tree; shape matching; multiresolution hierarchical description; coarse-to-fine shape information; symbolic description; range finder ClassCodes C1250 Article Type Theoretical / Mathematical Coden DTGDE7 Language Japanese RecordType Journal ControlNo. 4317321 AbstractNos. C9302-1250-173 References 13 Country Pub. Japan date 1204 ------------------------------------------------------------ Author Neumann, H.; Ottenberg, K.; Fachbereich Inf., Hamburg Univ., Germany Title Estimating attributes of smooth signal transitions from scale- space Source Proceedings. 11th IAPR International Conference on Pattern Recognition. Vol.III. Conference C: Image, Speech and Signal Analysis; Part: The Hague, Netherlands; Part: 30 Aug.-3 Sept. 1992; Sponsored by: Int. Assoc Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxv+791; 1992; pp. 754-8 Abstract Step-edge models as they have been used to model local intensity variation, only rarely are justified for the real case of image data. Due to finite apertures, the nature of scene geometry as well as discretization of the image, local intensity variations result in smooth transitions of varying width and local contrast. In order to appropriately deal with the robust detection and localization of image contrast, the authors propose the parametrized ramp transition as local signal model. The scale- space processing scheme for token extraction consists of a cascade of first band-pass filtering the raw data and a subsequent correlation of the result with a scaled first order derivative operator. The robust contrast detection within scale space and the estimation of local signal attributes in closed form is documented. The scheme can be extended to deal with intensity variations of different specificity Thesaurus edge detection; filtering and prediction theory Other Terms edge detection; smooth signal transitions; scale-space; parametrized ramp transition; local signal model; token extraction; first band-pass filtering; robust contrast detection ; signal attributes; intensity variations ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4311076 AbstractNos. B9302-6140C-129; C9302-1250-120 ISBN or SBN 0 8186 2920 7 References 12 Country Pub. USA date 1204 ------------------------------------------------------------ Author Kin, G.; Sato, M.; Precision & Intelligence Lab., Tokyo Inst. of Technol., Yokohama, Japan Title Scale space filtering on spherical pattern Source Proceedings. 11th IAPR International Conference on Pattern Recognition. Vol.III. Conference C: Image, Speech and Signal Analysis; Part: The Hague, Netherlands; Part: 30 Aug.-3 Sept. 1992; Sponsored by: Int. Assoc Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxv+791; 1992; pp. 638-41 Abstract For pattern information processing like recognition or understanding, one must express the given pattern properly for later processing. Many ways have been developed to express the hierarchy of a pattern, but many of them use one dimensional or two dimensional patterns. It is important to find an expression reflecting the hierarchical structure of a spherical pattern. In this paper, the authors propose the scale space filtering on a spherical pattern to express its hierarchy Thesaurus filtering and prediction theory; image recognition Other Terms 1D pattern; 2D pattern; Gaussian kernel; image recognition; convolution; spherical pattern; hierarchical structure; scale space filtering ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4311048 AbstractNos. B9302-6140C-106; C9302-1250-098 ISBN or SBN 0 8186 2920 7 References 3 Country Pub. USA date 1204 ------------------------------------------------------------ Author Neumann, H.; Ottenberg, K.; Stiehl, H.S.; Fachbereich Inf., Hamburg Univ., Germany Title Finding and describing local structure in discrete two- dimensional computed tomograms Source Proceedings. 11th IAPR International Conference on Pattern Recognition. Vol.III. Conference C: Image, Speech and Signal Analysis; Part: The Hague, Netherlands; Part: 30 Aug.-3 Sept. 1992; Sponsored by: Int. Assoc Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxv+791; 1992; pp. 408-12 Abstract Differently scaled intensity discontinuities along the gradient direction of scaled oriented contrast edges define significant X- ray CT and MR image structures. Reliable detection and quantitative description of such scaled intensity discontinuities is achieved with a scale-space approach. It results in an explicit token representation which can be understood as a compact symbolic description of significant image structure, namely a set of quantitative attributes per contrast edge. Moreover, a multi-resolution approach is used to reconstruct the image intensity function from the token representation alone using a membrane as prior image model. Current results from the novel entire processing cascade, which combines a quantitative discontinuity description and a membrane-based intensity reconstruction, are presented Thesaurus biomedical NMR; computerised tomography; edge detection; image reconstruction; medical diagnostic computing Other Terms 2D discrete computerised tomography; local structure description; edge detection; multiple resolution; image intensity function reconstruction; computed tomograms; X-ray CT; MR image structures; scaled intensity discontinuities; scale-space; explicit token representation; symbolic description; prior image model; membrane-based intensity reconstruction ClassCodes A8760G; A8760J; A8770E; C7330; C5260B; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4310991 AbstractNos. A9303-8760G-009; C9302-7330-021 ISBN or SBN 0 8186 2920 7 References 14 Country Pub. USA date 1204 ------------------------------------------------------------ Author van den Boomgaard, R.; Smeulders, A.W.M.; Fac. of Math. & Comput. Sci., Amsterdam Univ., Netherlands Title The morphological structure of images Source Proceedings. 11th IAPR International Conference on Pattern Recognition. Vol.III. Conference C: Image, Speech and Signal Analysis; Part: The Hague, Netherlands; Part: 30 Aug.-3 Sept. 1992; Sponsored by: Int. Assoc Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxv+791; 1992; pp. 268-71 Abstract The authors investigate the use of mathematical morphology to construct scale-spaces. These scale-spaces are based on differential equations, which are solved by morphological operators, describing the evolution of images in scale-space Thesaurus differential equations; image processing; mathematical morphology Other Terms image processing; morphological structure; mathematical morphology; differential equations; scale-space ClassCodes B6140C; B0220; C1250; C1120 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4304622 AbstractNos. B9301-6140C-360; C9301-1250-322 ISBN or SBN 0 8186 2920 7 References 10 Country Pub. USA date 1204 ------------------------------------------------------------ Author Jackway, P.T.; Centre for Signal Process. Res., Queensland Univ. of Technol., Brisbane, Qld., Australia Title Morphological scale-space Source Proceedings. 11th IAPR International Conference on Pattern Recognition. Vol.III. Conference C: Image, Speech and Signal Analysis; Part: The Hague, Netherlands; Part: 30 Aug.-3 Sept. 1992; Sponsored by: Int. Assoc Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxv+791; 1992; pp. 252-5 Abstract Scale-space is an important recent concept used in image processing and pattern recognition. Traditional scale-space is generated by a linear smoothing operation. The author presents a nonlinear type of smoother related to mathematical morphology which meets (modified) 'scale-space axioms' and also generates a 'scale-space'. This scale-space is full-plane and preserves the positions of features Thesaurus filtering and prediction theory; image processing; image recognition; mathematical morphology Other Terms nonlinear smoothing; morphological scale space; image processing ; pattern recognition; mathematical morphology ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4304618 AbstractNos. B9301-6140C-356; C9301-1250-318 ISBN or SBN 0 8186 2920 7 References 10 Country Pub. USA date 1204 ------------------------------------------------------------ Author Fermuller, C.; Kropatsch, W.; Center for Autom. Res., Maryland Univ., College Park, MD, USA Title Hierarchical curve representation Source Proceedings. 11th IAPR International Conference on Pattern Recognition. Vol.III. Conference C: Image, Speech and Signal Analysis; Part: The Hague, Netherlands; Part: 30 Aug.-3 Sept. 1992; Sponsored by: Int. Assoc Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxv+791; 1992; pp. 143-6 Abstract Presents a robust method for describing planar curves in multiple resolution using curvature information. The method is developed by taking into account the discrete nature of digital images as well as the discrete aspect of a multiresolution structure (pyramid). The authors deal with the robustness of the technique, which is due to the additional information that is extracted from observing the behavior of corners in the pyramid. Furthermore the resulting algorithm is conceptually simple and easily parallelizable. They develop theoretical results, analyzing the curvature of continuous curves in scale-space, which show the behavior of curvature extrema under varying scale. These results are used to eliminate any ambiguities that might arise from sampling problems due to the discreteness of the representation. Finally, experimental results demonstrate the potential of the method Thesaurus feature extraction; image processing; image recognition Other Terms hierarchical curve representation; image recognition; feature extraction; planar curves; digital images; multiresolution structure; pyramid; curvature; continuous curves ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4304592 AbstractNos. B9301-6140C-335; C9301-1250-299 ISBN or SBN 0 8186 2920 7 References 4 Country Pub. USA date 1204 ------------------------------------------------------------ Author Kok, T.C.W.; Dept. of Electron. Eng., City Polytech. of Hong Kong, Kowloon Tong, Hong Kong Title Multiresolution image segmentation Source Computer, Communication and Networking Systems: An Integrated Perspective. Proceedings of the International Conference on Information Engineering - ICIE '91; Part: Singapore; Part: 2-5 Dec. 1991; Amsterdam, Netherlands; Elsevier; 2 vol. xvi+1008; 1992; pp. 65-70 vol.1 Editor Subramanian, K.R.; Seumahu, E.S. Abstract This paper presents a novel segmentation technique based on dynamic thresholding using multiresolution images. A threshold surface is constructed by local segmentation across the multiresolution images formed by scale space filtering to adapt itself to learn the varying illumination from the image. Computer simulation is presented, and it shows promising results of the proposed technique Thesaurus digital simulation; image processing Other Terms segmentation technique; dynamic thresholding; multiresolution images; scale space filtering; illumination ClassCodes B6140C; C5260B; C6185 Article Type Practical Language English RecordType Conference ControlNo. 4292148 AbstractNos. B9301-6140C-097; C9301-5260B-074 ISBN or SBN 0 444 89480 2 References 10 Country Pub. Netherlands date 1195 ------------------------------------------------------------ ------------------------------------------------------------ Author Rangarajan, K.; Allen, W.; Shah, M.; Dept. of Comput. Sci., Central Florida, Orlando, FL, USA Title Recognition using motion and shape Source Proceedings. 11th IAPR International Conference on Pattern Recognition. Vol.1. Conference A: Computer Vision and Applications ; Part: The Hague, Netherlands; Part: 30 Aug.-3 Sept. 1992; Sponsored by: Int. Assoc. Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxv+795; 1992; pp. 255-8 Abstract Presents a method for matching sets of trajectories which supplements motion information with knowledge about the spatial relationships between points on the moving object. First the authors present a simple algorithm which matches two single trajectories using only motion information. They convert the 2D motion trajectories into two 1D signals based on the speed and direction components. The signals are then represented by scale- space images both to simplify matching and because the scale- space representations are translation and rotation invariant. They extend the matching algorithm to include spatial information and propose a second algorithm which matches multiple trajectories by combining motion and spatial match scores. Both algorithms were tested with real and synthetic data Thesaurus pattern recognition Other Terms 2D trajectory matching; shape information; pattern recognition; motion information; 2D motion trajectories; scale-space images; spatial information ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4286652 AbstractNos. B9301-6140C-060; C9301-1250-041 ISBN or SBN 0 8186 2910 X References 9 Country Pub. USA date 1204 ------------------------------------------------------------ Author Morita, S.; Kawashima, T.; Aoki, Y.; Fac. of Eng., Hokkaido Univ., Sapporo, Japan Title Hierarchical shape recognition based on 3-D multiresolution analysis Source Computer Vision - ECCV '92. Second European Conference on Computer Vision Proceedings; Part: Santa Margherita Ligure, Italy ; Part: 18-23 May 1992; Berlin, Germany; Springer-Verlag; xv+909; 1992; pp. 843-51 Editor Sandini, G. Abstract The paper introduces a method to create a hierarchical description of smooth curved surfaces based on scale-space analysis. The authors extend the scale-space method used in 1-D signal analysis to 3-D object. A 3-D scale-space images are segmented by zero-crossings of surface curvatures at each scale and then linked between consecutive scales based on topological changes (KH-description). The KH-description is, then, parsed and translated into the PS-tree which contains the number and distribution of subregions required for shape matching. The KH- description contains coarse-to-fine shape information of the object and the PS-tree is suitable for shape matching. A hierarchical matching algorithm using the descriptions is proposed and examples show that the symbolic description is suitable for efficient coarse-to-fine 3-D shape matching Thesaurus computer vision; image processing; pattern recognition Other Terms 3D objects; 3D multiresolution analysis; scale-space filtering; smooth curved surfaces; scale-space analysis; surface curvatures; topological changes; KH-description; PS-tree; shape matching; hierarchical matching algorithm; symbolic description ClassCodes B6140C; C5260B; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4286538 AbstractNos. B9301-6140C-051; C9301-5260B-041 ISBN or SBN 3 540 55426 2 References 10 Country Pub. Germany date 1201 ------------------------------------------------------------ Author Brunnstrom, K.; Lindeberg, T.; Eklundh, J.-O.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title Active detection and classification of junctions by foveation with a head-eye system guided by the scale-space primal sketch Source Computer Vision - ECCV '92. Second European Conference on Computer Vision Proceedings; Part: Santa Margherita Ligure, Italy ; Part: 18-23 May 1992; Berlin, Germany; Springer-Verlag; xv+909; 1992; pp. 701-9 Editor Sandini, G. Abstract The authors consider how junction detection and classification can be performed in an active visual system. This is to demonstrate that feature detection and classification in general can be done by both simple and robust methods, if the vision system is allowed to look at the world rather than at prerecorded images. The authors address the issue of how to attract attention to salient local image structures, as well as how to characterize them Thesaurus computer vision; pattern recognition Other Terms region of interest determination; foveation; head-eye system; scale-space primal sketch; junction detection; active visual system; feature detection; local image structures ClassCodes B6140C; C5260B; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4286518 AbstractNos. B9301-6140C-036; C9301-5260B-027 ISBN or SBN 3 540 55426 2 References 19 Country Pub. Germany date 1201 ------------------------------------------------------------ Author Florack, L.M.J.; ter Haar Romeny, B.M.; Koenderink, J.J.; Viergever, M.A.; Univ. Hospital, Utrecht, Netherlands Title Families of tuned scale-space kernels Source Computer Vision - ECCV '92. Second European Conference on Computer Vision Proceedings; Part: Santa Margherita Ligure, Italy ; Part: 18-23 May 1992; Berlin, Germany; Springer-Verlag; xv+909; 1992; pp. 19-23 Editor Sandini, G. Abstract The authors propose a formalism for deriving parametrised ensembles of local neighbourhood operators on the basis of a complete family of scale-space kernels, which are apt for the measurement of a specific physical observable. The parameters are introduced in order to associate a continuum of a priori equivalent kernels with each scale-space kernel, each of which is tuned to a particular parameter value. Ensemble averages, or other functional operations in parameter space, may provide robust information about the physical observable of interest. The approach gives a possible handle on incorporating multi- valuedness (transparency) and visual coherence into a single model. The authors consider the case of velocity tuning to illustrate the method. The emphasis, however, is on the formalism, which is more generally applicable Thesaurus computer vision; digital filters; filtering and prediction theory; picture processing Other Terms filter tuning; local neighbourhood operators; scale-space kernels; functional operations; parameter space; physical observable; multi-valuedness; transparency; visual coherence; velocity tuning ClassCodes B6140C; C5260B; C1250; C5240 Article Type Practical Language English RecordType Conference ControlNo. 4281795 AbstractNos. B9212-6140C-225; C9212-5260B-133 ISBN or SBN 3 540 55426 2 References 18 Country Pub. Germany date 1201 ------------------------------------------------------------ Author Toth, C.K.; Schenk, T.; Dept. of Geodetic Sci. & Surveying, Ohio State Univ., Columbus, OH, USA Title Feature-based matching for automatic image registration Source ITC Journal; ITC J. (Netherlands); no.1; 1992; pp. 40-6 Abstract For merging information extracted from satellite images with a GIS, the images are usually registered on a map by manually identifying a number of points on the satellite image, whose coordinates are measured on the map. The authors describe a method to solve the registration problem automatically. First, the authors develop a general scheme to extract features (edges) in the image and to match them with the corresponding features on the map. The authors take advantage of the multi-spectral resolution of the satellite image to perform the matching with selective features, for example with water bodies, cultural features or roads. A scale space approach is chosen for solving the problem of vastly different image scales. The paper concludes with experimental results Thesaurus computerised pattern recognition; computerised picture processing ; geographic information systems; remote sensing Other Terms feature extraction; edge extraction; computerised pattern recognition; remote sensing; feature-based mapping; computerised picture processing; automatic image registration; satellite images; GIS; multi-spectral resolution; water bodies; cultural features; roads; scale space approach ClassCodes C7840; C5260B Article Type Practical; Experimental Coden ITCJDP Language English RecordType Journal ControlNo. 4277205 AbstractNos. C9212-7840-079 ISSN 03032434 References 7 Country Pub. Netherlands date 1197 ------------------------------------------------------------ Author Eggert, D.W.; Bowyer, K.W.; Dyer, C.R.; Christensen, H.I.; Goldgof, D.B.; Dept. of Comput. Sci. & Eng., Univ. of South Florida, Tampa, FL, USA Title Applying the scale space concept to perspective projection aspect graphs Source Theory and Applications of Image Analysis. Selected Papers from the 7th Scandinavian Conference; Part: Aalborg, Denmark; Part: 13-16 Aug. 1991; Singapore; World Scientific; xiii+346; 1992; pp. 48-62 Editor Johansen, P.; Olsen, S. Abstract This paper reviews a complete implementation of an algorithm to compute the exact aspect graph of solids of revolution under perspective projection in 3-D space. The authors explore the notion of introducing scale into the qualitative aspect graph framework, thus providing a mechanism for selecting a level of detail that is 'large enough' to merit explicit representation. Several alternative interpretations of the scale space aspect graph are examined in response to the results produced for an example object by the implemented system Thesaurus graph theory; picture processing Other Terms perspective projection aspect graphs; algorithm; solids of revolution; explicit representation; scale space aspect graph ClassCodes B6140C; B0250; C1250; C1160 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4270048 AbstractNos. B9212-6140C-105; C9212-1250-078 ISBN or SBN 981 02 0945 2 References 36 Country Pub. Singapore date 1191 ------------------------------------------------------------ Author Lindeberg, T.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title On the behaviour in scale-space of local extrema and blobs Source Theory and Applications of Image Analysis. Selected Papers from the 7th Scandinavian Conference; Part: Aalborg, Denmark; Part: 13-16 Aug. 1991; Singapore; World Scientific; xiii+346; 1992; pp. 38-47 Editor Johansen, P.; Olsen, S. Abstract Elementary techniques from real analysis and singularity theory are applied to derive analytical results for the behaviour in scale-space of critical points and related entities. The main results of the treatment comprise a description of the general nature of trajectories of critical points in scale-space, an estimation of the drift velocity of critical points and edges, an analysis of the qualitative behaviour of critical points in bifurcation situations, and a classification of types of blob bifurcations possible Thesaurus critical points; picture processing Other Terms image analysis; scale-space; local extrema; singularity theory; critical points; trajectories; drift velocity; blob bifurcations ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4270047 AbstractNos. B9212-6140C-104; C9212-1250-077 ISBN or SBN 981 02 0945 2 References 17 Country Pub. Singapore date 1191 ------------------------------------------------------------ Author ter Haar Romeny, B.M.; Florack, L.M.J.; Koenderink, J.J.; Viergever, M.A.; Comput. Vision Res. Group, Utrecht Univ. Hospital, Netherlands Title Invariant third order properties of isophotes: T-junction detection Source Theory and Applications of Image Analysis. Selected Papers from the 7th Scandinavian Conference; Part: Aalborg, Denmark; Part: 13-16 Aug. 1991; Singapore; World Scientific; xiii+346; 1992; pp. 30-7 Editor Johansen, P.; Olsen, S. Abstract Geometric properties of isophotes are essential elements in image analysis techniques, due to their invariance under general invertible intensity transformations. The authors consider geometric properties which in addition are to be invariant under the group of rotations in the image domain. The third order local properties of isophotes are studied by a scale space approach. Numerical differentiation is replaced by convolution with Gaussian derivatives, to study higher order geometrical properties and to make a robust computer implementation. Isophotes at different intensity levels show a large change in curvature at T-junctions over a relatively small spatial scale, so that the gradient of isophote curvature is a good candidate for a T-junction detector Thesaurus picture processing Other Terms isophotes; image analysis; invertible intensity transformations; third order local properties; scale space approach; convolution; curvature; T-junction detector ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4270046 AbstractNos. B9212-6140C-103; C9212-1250-076 ISBN or SBN 981 02 0945 2 References 15 Country Pub. Singapore date 1191 ------------------------------------------------------------ Author Kawamata, M.; Kanbara, H.; Higuchi, T.; Dept. of Electron. Eng., Fac. of Eng., Tohoku Univ., Sendai, Japan Title Determination of IFS codes using scale-space correlation functions (iterated function systems) Source Workshop Notes. 1992 IEEE International Workshop on Intelligent Signal Processing and Communication Systems; Part: Taipei, Taiwan ; Part: 19-21 March 1992; Sponsored by: IEEE; Taipei, Taiwan; Nat. Taiwan Univ; xvii+603; 1992; pp. 219-33 Abstract Iterated function systems are regarded as time-variant state- space digital filters. IFS codes are obtained by determination of the numbers of contractive affine transformations and of their coefficients. Scale-space autocorrelation functions and coefficients are defined to determine coefficients of each contractive affine transformation. The important properties of these functions are proved for the determination of IFS codes. Coefficients of the contractive affine transformations can be determined from coordinates which maximize scale-space autocorrelation functions and coefficients of fractal images. Scale-space filtering is introduced in order to determine the number of affine transformations. First order zero-crossings are used to obtain a hierarchic description of scale-space images, which is a sequence of scale-space filtered images. This paper proposes an algorithm to determine IFS codes, and gives an illustrative example Thesaurus codes; correlation methods; data compression; digital filters; filtering and prediction theory; fractals; state-space methods Other Terms iterated function system codes; scale-space correlation functions ; time-variant state-space digital filters; contractive affine transformations; fractal images; zero-crossings; hierarchic description; scale-space filtered images; algorithm ClassCodes B6140C; B6120B; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4260247 AbstractNos. B9211-6140C-228; C9211-1250-178 References 6 Country Pub. Taiwan date 1199 ------------------------------------------------------------ Author Lindeberg, T.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title Scale-space behaviour of local extrema and blobs Source Journal of Mathematical Imaging and Vision; J. Math. Imaging Vis. (Netherlands); vol.1, no.1; March 1992; pp. 65-99 Abstract Elementary techniques from real analysis, and singularity theory, are applied to derive analytical results for the behaviour in scale-space of critical points and related entities. The main results of the treatment comprise a description of the general nature of trajectories of critical points in scale-space, estimates of the drift velocity of critical points and straight edges, an analysis of the qualitative behaviour of critical points in bifurcation situations, and a classification of what types of blob events are possible Thesaurus picture processing Other Terms scale-space behaviour; local extrema; blobs; real analysis; singularity theory; critical points; trajectories; drift velocity; straight edges; bifurcation situations ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Journal ControlNo. 4258520 AbstractNos. B9211-6140C-205; C9211-1250-164 ISSN 09249907 References 31 Country Pub. Netherlands date 1199 ------------------------------------------------------------ Author Mokhtarian, F.; Mackworth, A.K.; NTT Basic Res. Lab., Tokyo, Japan Title A theory of multiscale, curvature-based shape representation for planar curves Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.14, no.8; Aug. 1992; pp. 789-805 Abstract A shape representation technique suitable for tasks that call for recognition of a noisy curve of arbitrary shape at an arbitrary scale or orientation is presented. The method rests on the describing a curve at varying levels of detail using features that are invariant with respect to transformations that do not change the shape of the curve. Three different ways of computing the representation are described. They result in three different representations: the curvature scale space image, the renormalized curvature scale space image, and the resampled curvature scale space image. The process of describing a curve at increasing levels of abstraction is referred to as the evolution or arc length evolution of that curve. Several evolution and arc length evolution properties of planar curves are discussed Thesaurus pattern recognition; picture processing Other Terms multiscale curvature based shape representation; pattern recognition; image processing; planar curves; curvature scale space image; renormalized curvature scale space image; resampled curvature scale space image; arc length evolution ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4246272 AbstractNos. B9211-6140C-069; C9211-1250-039 ISSN 01628828 References 39 U.S. Copyright Clearance Center Code 0162-8828/92/$03.00 Country Pub. USA date 1204 ------------------------------------------------------------ Author Freeman, M.O.; Duell, K.A.; Fedor, A.; Nat. Sci. Found. Center for Optoelectron. Comput. Syst., Colorado Univ., Boulder, CO, USA Title Multi-scale optical image processing Source 1991 IEEE International Sympoisum on Circuits and Systems (Cat. No.91CH3006-4); Part: Singapore; Part: 11-14 June 1991; Sponsored by: IEEE; New York, NY, USA; IEEE; 5 vol. xlviii+3177; 1991; pp. 2355-8 vol.4 Abstract The authors introduce two multiscale optical image processing systems, a centroid scale-space processor and an optical wavelet processor. A.P. Witkin (1984) introduced the idea of scale-space, where information is presented on a coordinate system with continuous spatial and scale axes. His representation consists of a convolving the input signal with a Laplacian of a Gaussian point spread function and retaining the zero-crossings (which correspond to edges) as scale (width of the Gaussian) is varied. These edge maps form continuous contours in scale-space. The optical implementation of the system is discussed. The centroid scale-space map (CSSM) is constructed. This technique is extended to 2D input functions by using correlation techniques. The optical system which computes a 2D wavelet transform is shown. One special advantage is that the system complexity is independent of the scale parameter. This gives the flexibility to tailor the scale parameter to the particular application Thesaurus optical information processing; picture processing Other Terms spatial axes; multiscale optical image processing systems; centroid scale-space processor; optical wavelet processor; scale-space; coordinate system; scale axes; Gaussian point spread function; zero-crossings; edge maps; continuous contours ; optical implementation; correlation techniques; 2D wavelet transform; system complexity ClassCodes B6140C; B4180; C5270; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4244516 AbstractNos. B9211-6140C-032; C9211-5270-005 ISBN or SBN 0 7803 0050 5 References 11 U.S. Copyright Clearance Center Code CH3006-4/91/0000-2355$01.00 Country Pub. USA date 1189 ------------------------------------------------------------ Author Hattori, T.; Yamasaki, T.; Watanabe, Y.; Sanada, H.; Tezuka, Y.; Inf. & Comput. Sci. Lab., Kagawa Univ., Takamatsu, Japan Title Distance based vector field method for feature extraction of characters and figures Source Conference Proceedings 1991 IEEE International Conference on Systems, Man, and Cybernetics. 'Decision Aiding for Complex Systems (Cat. No.91CH3067-6); Part: Charlottesville, VA, USA; Par t: 13-16 Oct. 1991; Sponsored by: IEEE; New York, NY, USA; IEEE; 3 vol. xvi+2120; 1991; pp. 207-12 vol.1 Abstract A new method is proposed for feature extraction and representation of character and figure patterns using a vector field. The vector field is constructed from a distance transformation like the gradient vector of the electric field caused by a charged pattern. Features are defined by a set of the source points, sink points, and locally homogeneous domains of the vector field. It is shown that these features are effective for detecting holes (loops) and directional segments of the input pattern. The pattern and the frame can be segmented naturally and the segmentation represents the juxtapositional relation among the locally homogeneous domains of flow-out vectors from the pattern. The authors discuss the relation between the distance transformation that the vector field is based on and scale-space filtering which can be regarded as a kind of diffusion and/or blurring processing of the shape of pattern Thesaurus character recognition; filtering and prediction theory; pattern recognition Other Terms character recognition; shape analysis; pattern recognition; feature extraction; figure patterns; vector field; distance transformation; segmentation; juxtapositional relation; scale- space filtering ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4234255 AbstractNos. B9210-6140C-174; C9210-1250-138 ISBN or SBN 0 7803 0233 8 References 8 U.S. Copyright Clearance Center Code 0 7803 0233 8/91$01.00 Country Pub. USA date 1193 ------------------------------------------------------------ Author Fedor, A.; Freeman, M.O.; Optoelectron. Comput. Syst. Center, Colorado Univ., Boulder, CO, USA Title Optical multiscale morphological processor using a complex-valued kernel Source Applied Optics; Appl. Opt. (USA); vol.31, no.20; 10 July 1992; pp. 4042-50 Abstract Morphological transformations are typically performed on binary images by convolution with a binary kernel, which is followed by a threshold. The authors present an alternate approach that uses a complex-valued kernel with odd symmetry to perform these morphological operations. The complex-valued kernel increases the information-processing ability of the processor with no increase in system complexity. One advantage is that the processor operates on all constant regions of a gray-level image in parallel. A scale-space representation of this processor is obtained by varying the size of the kernel continuously through a range of scales. By using redundant information in the scale representation, this system is found to be robust in the presence of noise and spatial nonuniformities in the image. An optical system to perform morphological filtering based on this system is presented Thesaurus optical systems; picture processing; spatial filters Other Terms optical image noise; image processing; multiscale morphological processor; complex-valued kernel; binary images; convolution; binary kernel; threshold; odd symmetry; information-processing ability; system complexity; gray-level image; scale-space representation; scale representation; spatial nonuniformities; optical system; morphological filtering ClassCodes A4230V; B6140C Article Type Theoretical / Mathematical; Experimental Coden APOPAI Language English RecordType Journal ControlNo. 4230830 AbstractNos. A9220-4230-003; B9210-6140C-154 ISSN 00036935 References 23 U.S. Copyright Clearance Center Code 0003-6935/92/204042-09$05.00/0 Country Pub. USA date 1203 ------------------------------------------------------------ Author Florack, L.M.J.; ter Haar Romeny, B.M.; Koenderink, J.J.; Viergever, M.A.; Comput. Vision Res. Group, Utrecht Univ. Hospital, Netherlands Title Scale and the differential structure of images Source Image and Vision Computing; Image Vis. Comput. (UK); vol.10, no.6; July-Aug. 1992; pp. 376-88 Abstract Scaled partial differential operators enable local image analysis through the detection of local differential structure in a robust way, while at the same time capturing global features through the extra scale degree of freedom. The paper shows why the operations of scaling and differentiation cannot be separated. This framework permits one to construct in a systematic way multiscale, cartesian differential invariants, i.e. true image descriptors that exhibit manifest invariance with respect to a change of cartesian coordinates. The scale-space operators closely resemble the receptive field profiles found in mammalian front-end visual systems Thesaurus computerised picture processing Other Terms scaled partial differential operators; differential structure; local image analysis; local differential structure; global features; scaling; differentiation; cartesian differential invariants; true image descriptors; manifest invariance; cartesian coordinates; scale-space operators; receptive field profiles; mammalian front-end visual systems ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Coden IVCODK Language English RecordType Journal ControlNo. 4219231 AbstractNos. B9210-6140C-024; C9210-1250-019 ISSN 02628856 References 31 U.S. Copyright Clearance Center Code 0262-8856/92/006376-13 Country Pub. UK date 1203 ------------------------------------------------------------ Author Goshtasby, A.; Dept. of Electr. Eng. & Comput. Sci., Illinois Univ., Chicago, IL, USA Title Gaussian decomposition of two-dimensional shapes: a unified representation for CAD and vision applications Source Pattern Recognition; Pattern Recognit. (UK); vol.25, no.5; May 1992; pp. 463-72 Abstract A new parametric representation for two-dimensional (2D) shapes is introduced using Gaussians as basis functions. In this representation, to design a shape, the parameters of the Gaussians are specified. To recognize a shape, the shape is decomposed into Gaussians and the parameters of the Gaussians are used as features. Methods to design free-form shapes in CAD applications and recognize shapes in vision applications are described using the proposed representation Thesaurus CAD; computer graphics; computer vision; computerised pattern recognition Other Terms 2D shapes; Gaussian decomposition; computer vision; scale space images; CAD; parametric representation ClassCodes B6140C; C1250; C5260B; C6130B; C7400 Article Type Applications; New Development; Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 4197957 AbstractNos. B9209-6140C-015; C9209-1250-020 ISSN 00313203 References 51 U.S. Copyright Clearance Center Code 0031-3203/92/$5.00+.00 Country Pub. UK date 1201 ------------------------------------------------------------ Author Guo-Li Ao; Yu-Jun Cui; Izumi, M.; Fukunaga, K.; Shanghai Teacher's Univ., China Title Structural tree representation of outlines and its application to object recognition Source Transactions of the Institute of Electronics, Information and Communication Engineers D-II; Trans. Inst. Electron. Inf. Commun. Eng. D-II (Japan); vol.J75D-II, no.3; March 1992; pp. 481-9 Abstract Scale-space filtering has been discussed in order to represent the waveforms in hierarchical forms. But in a special condition there may be large difference between the shapes of its zero crossing lines of second derivative even if the different between the structures of waveforms is not large. In addition, in the description of the second zero crossing lines by graphs (tree structure) and vice versa. It is difficult to describe the structure by tree structure of previous manner. To solve these problems, the authors introduce an equivalent transformation between tree structures and propose the similarity between tree structures. It is shown that the matching between outline figures of objects can be efficiently examined, and an application of the scale space approach to object recognition is discussed Thesaurus pattern recognition; picture processing; trees (mathematics) Other Terms waveforms; hierarchical forms; zero crossing lines; second derivative; graphs; tree structure; scale space approach; object recognition ClassCodes C1250; C1160 Article Type Practical Coden DTGDE7 Language Japanese RecordType Journal ControlNo. 4192321 AbstractNos. C9208-1250-295 References 9 Country Pub. Japan date 1199 ------------------------------------------------------------ Author Jeong, H.; Kim, C.I.; Dept. of Electr. Eng., Pohang Inst. of Sci. & Technol., South Korea Title Adaptive determination of filter scales for edge detection Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.14, no.5; May 1992; pp. 579-85 Abstract The authors suggest a regularization method for determining scales for edge detection adaptively for each site in the image plane. Specifically, they extend the optimal filter concept of T. Poggio et al. (1984) and the scale-space concept of A. Witkin (1983) to an adaptive scale parameter. To avoid an ill-posed feature synthesis problem, the scheme automatically finds optimal scales adaptively for each pixel before detecting final edge maps. The authors introduce an energy function defined as a functional over continuous scale space. Natural constraints for edge detection are incorporated into the energy function. To obtain a set of optimal scales that can minimize the energy function, a parallel relaxation algorithm is introduced. Experiments for synthetic and natural scenes show the advantages of the algorithm. In particular, it is shown that this system can detect both step and diffuse edges while drastically filtering out the random noise Thesaurus filtering and prediction theory; pattern recognition; picture processing Other Terms adaptive scale determination; filter scales; edge detection; regularization method; optimal filter; adaptive scale parameter; energy function; continuous scale space; parallel relaxation algorithm ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4185381 AbstractNos. B9208-6140C-226; C9208-1250-228 ISSN 01628828 References 24 U.S. Copyright Clearance Center Code 0162-8828/92/$03.00 Country Pub. USA date 1201 ------------------------------------------------------------ Author Grosky, W.I.; Jiang, Z.; Dept. of Comput. Sci., Wayne State Univ., Detroit, MI, USA Title A hierarchical approach to feature indexing Source Image Storage and Retrieval Systems; Part: San Jose, CA, USA; Par t: 13-14 Feb. 1992; Sponsored by: SPIE; IS&T; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1662; 1992; pp. 9-20 Abstract The authors extend their previous approach to 2D shape indexing, data-driven indexed hypotheses, to hierarchical features. It is shown mathematically and experimentally that an index based on hierarchical features is more computationally efficient than one based on non-hierarchical features. Their approach addresses two types of hierarchies: a multilevel approximation of the contours of 2D objects and a three-level feature indexing system. The approach can easily be extended to 3D objects. Occluded object recognition should be based on local features. However, these features may sometimes be lost due to changes of scale as well as to occlusion. The first hierarchical mechanism is used to complement the feature loss due to scale changes. It results in multilevel approximations of the contours of objects using scale- space approaches. This approach will also be beneficial when there are few boundary points of maximal curvature so that standard polygonal approximation schemes don't work very well. The second hierarchical mechanism is to use sets of visible local features to hypothesize the presence of objects. Verification of the various hypotheses is done via normalization and boundary template matching Thesaurus computerised pattern recognition; database management systems; indexing; information retrieval Other Terms image databases; occluded object recognition; visual databases; feature indexing; 2D shape indexing; data-driven indexed hypotheses; hierarchical features; multilevel approximation; contours; three-level feature indexing; feature loss; scale changes; scale-space; visible local features; normalization; boundary template matching ClassCodes C6160S; C5260B Article Type Practical; Experimental Coden PSISDG Language English RecordType Conference ControlNo. 4184535 AbstractNos. C9208-6160S-005 ISSN 0277786X References 18 U.S. Copyright Clearance Center Code 0 8194 0816 6/92/$4.00 Country Pub. USA date 1198 ------------------------------------------------------------ Author Radha, H.; Leonardi, R.; Vetterli, M.; AT&T Bell Lab., Holmdel, NJ, USA Title A multiresolution approach to binary tree representations of images Source ICASSP 91: 1991 International Conference on Acoustics, Speech and Signal Processing (Cat. No.91CH2977-7); Part: Toronto, Ont., Canada; Part: 14-17 April 1991; Sponsored by: IEEE; New York, NY, USA; IEEE; 5 vol. 3732; 1991; pp. 2653-6 vol.4 Abstract A multiresolution method for constructing a BSP (binary space partitioning) tree is introduced. This approach derives a hierarchy (pyramid) of scale-space images from the original image. In this hierarchy, a BSP tree of an image is built from other trees representing low-resolution images of the pyramid. A low- resolution image BSP tree serves as an initial guess to construct a higher-resolution image tree. Due to filtering when constructing the pyramid, details are discarded. As a result, a more robust segmentation is obtained. Moreover a significant computational advantage is achieved Thesaurus filtering and prediction theory; picture processing; trees (mathematics) Other Terms image segmentation; binary tree representations; multiresolution method; binary space partitioning; scale-space images; low-resolution images; higher-resolution image; filtering ClassCodes B6140C; B0250; C1250; C1160 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4183670 AbstractNos. B9208-6140C-185; C9208-1250-189 ISBN or SBN 0 7803 0003 3 References 7 U.S. Copyright Clearance Center Code CH2977-7/91/0000-2653$01.00 Country Pub. USA date 1187 ------------------------------------------------------------ Author Tai-Yuen Cheng; Chung-Lin Huang; Dept. of Electr. Eng., Nat. Tsing-Hua Univ., Hsin-Chu, Taiwan Title Color images segmentation using scale space filter and Markov random field Source Intelligent Robots and Computer Vision X: Algorithms and Techniques; Part: Boston, MA, USA; Part: 11-13 Nov. 1991; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1607; 1992; pp. 358-68 Abstract The paper presents a hybrid method that combines the scale space filter (SSF) and Markov random field (MRF) for color image segmentation. Using the scale space filter, the authors separate the different scaled histogram to intervals corresponding to peaks and valleys. The basic construction of MRF is a joint probability given the original data. The original data is the image that the authors obtained from the source and the result is called the label image. Because the MRF needs the number of segments before it converges to the global minimum, they exploit the scale space filter to do coarse segmentation and then use MRF to do fine segmentation of the images. Finally, they compare the experimental results obtained from using SSF only, or combined with MRF using iterated conditional mode (ICM), and Gibbs sampling Thesaurus colour; Markov processes; pattern recognition; picture processing Other Terms scale space filter; Markov random field; color image segmentation; joint probability; coarse segmentation; fine segmentation; iterated conditional mode; Gibbs sampling ClassCodes B6140C; C1250; C1260 Article Type Practical Coden PSISDG Language English RecordType Conference ControlNo. 4174950 AbstractNos. B9208-6140C-037; C9208-1250-036 ISSN 0277786X References 11 U.S. Copyright Clearance Center Code 0 8194 0744 5/92/$4.00 Country Pub. USA date 1194 ------------------------------------------------------------ Author Matsopoulos, G.; Marshall, S.; Strathclyde Univ., Glasgow, UK Title A new morphological scale space operator Source International Conference on Image Processing and its Applications (Conf. Publ. No.354); Part: Maastricht, Netherlands; Part: 7-9 April 1992; London, UK; IEE; xxiv+628; 1992; pp. 246-9 Abstract Considers the technique of scale space for a range of image processing and computer vision applications. In the past both linear and morphological approaches have been used to evaluate a scale space description and it has been shown that morphological techniques possess certain advantages including speed of computation. However the existing morphological approaches produce a non symmetrical output with respect to intrusions and protrusions of a signal. The paper describes a new morphological approach for scale space description which eliminates the above problem and which therefore treats both signal intrusions and protrusions equally. Also, a new property called the AVEC property, related to scale space behaviour is defined Thesaurus computer vision; computerised picture processing Other Terms image processing; computer vision; morphological approach; scale space; signal intrusions; protrusion; AVEC property; scale space behaviour ClassCodes C1250; C5260B Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4174157 AbstractNos. C9207-1250-219 ISBN or SBN 0 85296 543 5 References 5 Country Pub. UK date 1200 ------------------------------------------------------------ Author Forte, P.; Netherwood, P.; Barnwell, P.; Kingston Polytech., Kingston upon Thames, UK Title Spiral arc shape representation for automatic inspection of surface mount assemblies Source International Conference on Image Processing and its Applications (Conf. Publ. No.354); Part: Maastricht, Netherlands; Part: 7-9 April 1992; London, UK; IEE; xxiv+628; 1992; pp. 143-6 Abstract Describes an efficient technique for representing 2D shapes, by approximating their contours with segments of spiral arcs. By definition a spiral arc is one which exhibits monotone curvature along the arc length, straight line segments and circular arcs being special cases. The arc is specified by giving the position of each endpoint, plus the orientation of the tangent to the arc at each endpoint. For a straight line or circular arc this information specifies the curve uniquely. For other arcs a formula is given for generating an arc between the two points. This formula minimizes the curvature of the arc subject to various smoothness constraints. The generated arc may only approximate the true contour to within a specified level of tolerance, where the degree of tolerance constitutes a scale space parameter Thesaurus computerised picture processing Other Terms shape representation; automatic inspection; surface mount assemblies; 2D shapes; spiral arcs; monotone curvature; tolerance; scale space parameter ClassCodes C5260B Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4169499 AbstractNos. C9207-5260B-124 ISBN or SBN 0 85296 543 5 References 6 Country Pub. UK date 1200 ------------------------------------------------------------ Author Lu, Y.; Jain, R.C.; Environ. Res. Inst. of Michigan, Ann Arbor, MI, USA Title Reasoning about edges in scale space Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.14, no.4; April 1992; pp. 450-68 Abstract Explores the role of reasoning in early vision processing. In particular, the problem of detecting edges is addressed. The authors do not try to develop another edge detector, but rather, they study an edge detector rigorously to understand its behavior well enough to formulate a reasoning process that allow appliance of the detector judiciously to recover useful information. They present a multiscale reasoning algorithm for edge recovery: reasoning about edges in scale space (RESS). The knowledge in RESS is acquired from the theory of edge behavior in scale space and represented by a number of procedures. RESS recovers desired edge curves through a number of reasoning processes on zero crossing images at various scales. The knowledge of edge behavior in scale space enables RESS to select proper scale parameters, recover missing edges, eliminate noise or false edges, and correct the locations of edges. A brief evaluation of RESS is performed by comparing it with two well-known multistage edge detection algorithms Thesaurus computer vision; inference mechanisms; knowledge representation Other Terms computer vision; knowledge representation; inference mechanisms; noise elimination; edges; scale space; edge detector; multiscale reasoning algorithm; edge recovery; edge behavior; edge curves; zero crossing images ClassCodes C5260B; C6170 Article Type Practical; Theoretical / Mathematical; Experimental Coden ITPIDJ Language English RecordType Journal ControlNo. 4166328 AbstractNos. C9207-5260B-093 ISSN 01628828 References 38 U.S. Copyright Clearance Center Code 0162-8828/92/$03.00 Country Pub. USA date 1200 ------------------------------------------------------------ Author Rattarangsi, A.; Chin, R.T.; Dept. of Electr. & Comput. Eng., Wisconsin Univ., Madison, WI, USA Title Scale-based detection of corners of planar curves Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.14, no.4; April 1992; pp. 430-49 Abstract A technique for detecting and localizing corners of planar curves is proposed. The technique is based on Gaussian scale space, which consists of the maxima of absolute curvature of the boundary function presented at all scales. The scale space of isolated simple and double corners is first analyzed to investigate the behavior of scale space due to smoothing and interactions between two adjacent corners. The analysis shows that the resulting scale space contains line patterns that either persist, terminate, or merge with a neighboring line. Next, the scale space is transformed into a tree that provides simple but concise representation of corners at multiple scales. Finally, a multiple-scale corner detection scheme is developed using a coarse-to-fine tree parsing technique. The parsing scheme is based on a stability criterion that states that the presence of a corner must concur with a curvature maximum observable at a majority of scales. Experiments were performed to show that the scale space corner detector is reliable for objects with multiple- size features and noisy boundaries and compares favorably with other corner detectors tested Thesaurus filtering and prediction theory; pattern recognition; picture processing; trees (mathematics) Other Terms pattern recognition; picture processing; scale-based corners detection; planar curves; Gaussian scale space; maxima of absolute curvature; boundary function; line patterns; tree; multiple-scale corner detection; coarse-to-fine tree parsing technique; stability criterion ClassCodes B6140C; B0250; C1250; C1160 Article Type Theoretical / Mathematical; Experimental Coden ITPIDJ Language English RecordType Journal ControlNo. 4166327 AbstractNos. B9207-6140C-173; C9207-1250-150 ISSN 01628828 References 22 U.S. Copyright Clearance Center Code 0162-8828/92/$03.00 Country Pub. USA date 1200 ------------------------------------------------------------ Author Saund, E.; Dept. of Brain & Cognitive Sci., MIT, Cambridge, MA, USA Title Putting knowledge into a visual shape representation Source Artificial Intelligence; Artif. Intell. (Netherlands); vol.54, no.1-2; March 1992; pp. 71-119 Abstract Shows how a representation for visual shape can be formulated to employ knowledge about the geometrical structures common within specific shape domains. In order to support a wide variety of later visual processing tasks, the authors seek representations making explicit many geometric properties and spatial relationships redundantly and at many levels of abstraction. The authors offer two specific computational tools: (1) By maintaining shape tokens on a scale-space blackboard, information about the relative locations and sizes of shape fragments such as contours and regions can be manipulated symbolically, while the pictorial organization inherent to a shape's spatial geometry is preserved. (2) Through the device of dimensionality-reduction, configurations of shape tokens can be interpreted in terms of their membership within deformation classes; this provides leverage in distinguishing shapes on the basis of subtle variations reflecting deformations in their forms. Using these tools, knowledge in a shape representation resides in a vocabulary of shape descriptors naming constellations of shape tokens in the scale-space blackboard. The approach is illustrated through a computer implementation of a hierarchical shape vocabulary designed to offer flexibility in supporting important aspects of shape recognition and shape comparison in the two- dimensional shape domain of the dorsal fins of fishes Thesaurus artificial intelligence; knowledge representation; visual perception Other Terms visual shape representation; knowledge; geometrical structures; geometric properties; spatial relationships; computational tools; shape tokens; scale-space blackboard; shape fragments; pictorial organization; dimensionality-reduction; deformation classes; constellations; two-dimensional shape domain; dorsal fins ClassCodes C1230; C6170 Article Type Theoretical / Mathematical Coden AINTBB Language English RecordType Journal ControlNo. 4165157 AbstractNos. C9207-1230-042 ISSN 00043702 References 44 U.S. Copyright Clearance Center Code 0004-3702/92/$05.00 Country Pub. Netherlands date 1199 ------------------------------------------------------------ Author Topkar, V.A.; Sood, A.K.; Dept. of Comput. Sci., George Mason Univ., Fairfax, VA, USA Title Statistical analysis of scale-space (computer vision) Source Signal Processing; Signal Process. (Netherlands); vol.26, no.3; March 1992; pp. 307-34 Abstract The noise and clutter cause major problems in scale-space because it involves taking higher derivatives. Statistical analysis of the noise as it is reflected in the scale-space representation and its effect on any algorithm based on this representation poses an interesting topic of research. Scale-space representation involves convolving the input with a smoothing filter (typically Gaussian) of different resolutions and detecting the primitives (typically the zero crossings of the second derivative) to form the scale-space representation. Statistical analysis of the scale-space is nontrivial because of two reasons: (i) it involves a nonlinear operation, namely the detection of zero crossings and (ii) the noise at different scales is correlated. The authors prove theorems which give the probabilities of zero-crossings in the output in the presence of noise. The theorems are then applied to the case of Gaussian smoothing. These probabilities can be used for a number of applications such as performance analysis Thesaurus computer vision; filtering and prediction theory; picture processing; probability; statistical analysis Other Terms statistical analysis; zero-crossings probability; computer vision; noise; clutter; scale-space representation; smoothing filter; Gaussian smoothing ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden SPRODR Language English RecordType Journal ControlNo. 4159846 AbstractNos. B9207-6140C-081; C9207-1250-071 ISSN 01651684 References 35 U.S. Copyright Clearance Center Code 0165-1684/92/$05.00 Country Pub. Netherlands date 1199 ------------------------------------------------------------ Author Takamatsu, R.; Kimura, K.; Sato, M.; Kawarada, H.; Tokyo Inst. of Technol., Yokohama, Japan Title Pitch determination with scale space filtering Source Bulletin of Precision and Intelligence Laboratory; Bull. Precis. Intell. Lab. (Japan); no.66; Sept. 1991; pp. 1-7 Abstract In conventional methods of pitch determination based on short- term analysis, the pitch frequency fluctuates according to frame length and positioning. Furthermore, pitch frequency is averaged in the frame. This paper introduces a new pitch determination algorithm using scale space filtering. Scale-space filtering is a multi-resolutional filtering method which expands a waveform into a set of waveforms for multi-scale measurement. Its zero- crossings naturally describe the hierarchical structure of the waveform. In the method, frame length is decided by the structure of the waveform and pitch frequency is directly derived from a single pitch period. Therefore the fluctuations due to frame length and positioning are diminished and the changes of pitch frequency are accurately determined. The result of an experiment with the weather forecast sentence shows that the method performs quite well Thesaurus filtering and prediction theory; speech analysis and processing Other Terms scale space filtering; pitch determination; pitch frequency; frame length; multi-resolutional filtering method; zero- crossings; hierarchical structure; frame length; weather forecast sentence ClassCodes B6130; B6140 Article Type Theoretical / Mathematical Language English RecordType Journal ControlNo. 4154089 AbstractNos. B9206-6130-071 ISSN 03857832 References 3 Country Pub. Japan date 1192 ------------------------------------------------------------ Author Doi, S.; Mori, K.; Takahashi, H.; Shimizu, E.; Matsuda, M.; Nara Nat. Coll. of Technol., Yamatokoriyama, Japan Title Parallel processing of the scale-space filtering using the transputer Source Applications of Transputers 3. Proceedings of the Third International Conference on Applications of Transputers; Part: Glasgow, UK; Part: 28-30 Aug. 1991; Sponsored by: UK SERC/DTI Initiative on the Eng. Appl. Transputers; IEEE; IEE; IOP; et al; Amsterdam, Netherlands; IOS; 821; 1991; pp. 571-6 Editor Durrani, T.S.; Sandham, W.A.; Soraghan, J.J.; Forbes, S.M. Abstract Two types of parallel processing algorithms are proposed for the high speed calculation of scale-space filtering using the transputer. The parallel algorithm utilizing the scale axis division formula is executed by assigning the calculation divided by scale parameter to each transputer. In this algorithm, the measuring results show that the execution time for calculation has been reduced proportional to the number of transputers. In the parallel algorithm utilizing the adaptive search formula, each zero-cross point is found at the same time from the small area in which the existence of each zero-cross point is suspected Thesaurus computerised signal processing; filtering and prediction theory; parallel algorithms Other Terms parallel processing; scale-space filtering; transputer; parallel processing algorithms; scale axis division formula; adaptive search formula; zero-cross point ClassCodes B6140; C1260; C4240P Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4144457 AbstractNos. B9206-6140-209; C9206-1260-154 References 7 Country Pub. Netherlands date 1191 ------------------------------------------------------------ Author Janssen, H.; Kopecz, J.; Inst. for Neuroinformatics, Ruhr-Univ. of Bochum, Germany Title Image recognition in hypercolumnar scale space by sparsely coded associative memory Source Artificial Neural Networks. Proceedings of the 1991 International Conference. ICANN-91; Part: Espoo, Finland; Part: 24-28 June 1991; Amsterdam, Netherlands; North-Holland; 2 vol. xix+1819; 1991; pp. 1203-6 vol.2 Editor Kohonen, T.; Makisara, K.; Simula, O.; Kangas, J. Abstract The authors propose a pattern recognition system based on an architecture close to the one found in human visual cortex which is called hypercolumns. They show that this discrete parametric representation can be used to define a short range interaction to make desirable information like edge continuation more explicit. They also show how hypercolumnar representations can be sparsely coded for usage in a very efficient associative memory recognition system. They combine this system with a model for coarse-to-fine search in hypercolumnar scale space thus gaining translational invariance. In principle the application of such a representation appears to be very well suited for data reduction and pattern recognition processes and is part of a neural instruction set Thesaurus computerised pattern recognition; computerised picture processing ; content-addressable storage Other Terms hypercolumnar scale space; sparsely coded associative memory; pattern recognition; discrete parametric representation; short range interaction; coarse-to-fine search; translational invariance; data reduction; neural instruction set ClassCodes C5340; C5260B; C1250 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 4143967 AbstractNos. C9206-5340-023 ISBN or SBN 0 444 89178 1 References 8 Country Pub. Netherlands date 1189 ------------------------------------------------------------ Author Topkar, V.A.; Sood, A.K.; Kjell, B.; Dept. of Comput. Sci., George Mason Univ., Fairfax, VA, USA Title Object detection using contrast based scale-space Source Proceedings 1991 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (91CH2983-5); Part: Maui, HI, USA; Part: 3-6 June 1991; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Sco. Press; xvi+761; 1991; pp. 700-1 Abstract The authors propose four scale-space object detection algorithms for separating objects from the background. These algorithms do not need thresholding at any of the scales. The different algorithms are applicable to images with different noise and clutter characteristics. Statistical analysis of the four algorithms is conducted for noisy and cluttered backgrounds Thesaurus pattern recognition; picture processing Other Terms statistical analysis; edge focusing; contrast based scale-space; scale-space object detection; noise; clutter characteristics ClassCodes C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4137196 AbstractNos. C9206-1250-036 ISBN or SBN 0 8186 2148 6 References 3 U.S. Copyright Clearance Center Code CH2983-5/91/0000-0700$01.00 Country Pub. USA date 1189 ------------------------------------------------------------ Author Kulkarni, D.; Kutulakos, K.; Robinson, P.; NASA Ames Res. Center, Moffett Field, CA, USA Title Data analysis using scale-space filtering and Bayesian probabilistic reasoning (for differential thermal analysis) Source Computers & Chemistry; Comput. Chem. (UK); vol.16, no.1; Jan. 1992; pp. 15-23 Abstract Describes a program for the analysis of output curves from a differential thermal analyzer (DTA). The program first extracts probabilistic qualitative features from a DTA curve of a soil sample, and then uses Bayesian probabilistic reasoning to infer what minerals are present in the soil. It consists of a qualifier module and a classifier module. The qualifier employs a simple and efficient extension of scale-space filtering DTA data. Ordinarily, when filtering operations are not highly accurate, points can vanish from contours in the scale-space image. To handle the problem of vanishing points, the authors' algorithm uses perceptual organization heuristics to group the points into lines. It then groups these lines into contours by using additional heuristics. Probabilities are associated with these contours using domain-specific correlations. A Bayes tree classifier processes probabilistic features to infer the presence of different minerals in the soil. The authors show experimentally that using domain-specific correlations to infer qualitative features, this algorithm outperforms a domain- independent algorithm that does not Thesaurus Bayes methods; chemistry computing; classification; data analysis; filtering and prediction theory; geochemistry; heuristic programming; inference mechanisms; minerals; probability; soil; thermal analysis; trees (mathematics) Other Terms data analysis; output curves analysis; scale-space filtering; Bayesian probabilistic reasoning; differential thermal analyzer; probabilistic qualitative features; soil; minerals; qualifier module; classifier module; contours; vanishing points; perceptual organization heuristics; lines; domain-specific correlations; Bayes tree classifier ClassCodes A0650; A0250; C7320; C1140Z; C1260; C7340; C1230 Article Type Practical Coden COCHDK Language English RecordType Journal ControlNo. 4133272 AbstractNos. A9211-0650-001; C9206-7320-003 ISSN 00978485 References 12 U.S. Copyright Clearance Center Code 0097-8485/92/$5.00+0.00 Country Pub. UK date 1197 ------------------------------------------------------------ Author Jager, F.; Koren, I.; Gyergyek, L.; Fac. of Electr. & Comput. Eng., Ljubljana, Yugoslavia Title Multiresolution representation and analysis of ECG waveforms Source Proceedings. Computers in Cardiology (Cat. No.90CH3011-4); Part: Chicago, IL, USA; Part: 23-26 Sept. 1990; Sponsored by: IEEE; Illinois Inst. Technol.; Pritzker Inst. Med. Eng.; Nat. Inst. Health.; American Heart Assoc. of Metropolitan Chicago; et al; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xxii+711; 1991; pp. 547-50 Abstract A new pattern recognition method is introduced for the multiresolution representation and analysis of electrocardiogram (ECG) waveforms. The multiresolution representation is based on filtering the curvature of the curve with continuum of Gaussian filters where Gaussian standard deviation increases, and on extracting of extrema points in filtered versions of the curvature (scale-space filtering). The original curve is then segmented at each scale into linear parts with regard to the extracted extrema points. After segmentation and linking segments between scales, shape is represented qualitatively in a hierarchical tree form holding information on coarser and finer details of shape. Different methods of tree form analysis can be applied to data compression, classification of heart beats or fine structure analysis. The fast computation scheme and transformation into hierarchical structure are described. In addition to the method of representation, a data compression method is proposed. Comparison to the AZTEC data compression method is given Thesaurus electrocardiography; pattern recognition; waveform analysis Other Terms heart beats classification; curvature filtering; ECG waveforms analysis; pattern recognition method; multiresolution representation; Gaussian standard deviation; scale-space filtering; extrema points; hierarchical tree form; hierarchical structure; AZTEC data compression method ClassCodes A8770E; A8730C; A8728; B7510D Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4127861 AbstractNos. A9210-8770E-023; B9205-7510D-049 ISBN or SBN 0 8186 2225 3 References 6 U.S. Copyright Clearance Center Code 0276-6574/91/0000-0547$01.00 Country Pub. USA date 1179 ------------------------------------------------------------ Author Sato, J.; Sato, M.; Precision Intelligence Lab., Tokyo Inst. of Technol., Yokohama, Japan Title A local structural analysis of images based on scale space filtering Source Transactions of the Institute of Electronics, Information and Communication Engineers D-II; Trans. Inst. Electron. Inf. Commun. Eng. D-II (Japan); vol.J74D-II, no.12; Dec. 1991; pp. 1715-22 Abstract To realize artificial vision it is one of the important themes to establish a good representation of pattern images for recognition and understanding. We humans look at something hierarchically. We see it to know its rough structure at first, then we view it to understand its details. Scale space filtering is the operation of convoluting a pattern image with the Gaussian blurring function, which is an effective method to handle images hierarchically. In this study, the authors analyze how the figure of ridge-valley lines of the tone level of the image changes near the singular stationary points locally. As a result of the analysis, they find that the singular stationary points are classified into three basic types, that is hyperbolic type, positive elliptic type, and negative elliptic type Thesaurus computer vision Other Terms image representation; pattern recognition; pattern understanding ; convolution; hierarchical method; local structural analysis; scale space filtering; artificial vision; rough structure; details; Gaussian blurring function; ridge-valley lines; tone level; singular stationary points; hyperbolic type; positive elliptic type; negative elliptic type ClassCodes C1250 Article Type Theoretical / Mathematical Coden DTGDE7 Language Japanese RecordType Journal ControlNo. 4125406 AbstractNos. C9205-1250-060 References 7 Country Pub. Japan date 1195 ------------------------------------------------------------ Author Etoh, M.; Tomono, A.; Kishino, F.; ATR Commun. Syst. Res. Labs., Kyoto, Japan Title Stereo-based description by generalized cylinder complexes from occluding contours Source Systems and Computers in Japan; Syst. Comput. Jpn. (USA); vol.22, no.12; 1991; pp. 79-89 Abstract The paper proposes a stereo-based description technique as a preprocessing step for three-dimensional (3-D) recognition of an object representable by a straight homogeneous generalized cylinder (SHGC). The 3-D position estimate of the SHGC model be computed if the object being modeled is either a solid of revolution or a body whose contour is symmetrical with respect to its axial. The proposed algorithm is applicable to a complex object if that object can be partitioned into a set of multiple SHGCs. The most important task of the algorithm is the extraction of a symmetry axis for each GC section of the object. The contour lines are analyzed in a scale space. The analysis is performed recursively to obtain each CC section that has stable contour lines. The effectiveness of the proposed algorithm for 3-D description of an object has been demonstrated via experimentation Thesaurus pattern recognition; picture processing Other Terms 3D object recognition; stereo imaging; generalized cylinder complexes; occluding contours; stereo-based description; solid of revolution; symmetry axis; GC section; contour lines; stable contour lines ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden SCJAEP Language English RecordType Journal ControlNo. 4115131 AbstractNos. B9205-6140C-032; C9205-1250-023 ISSN 08821666 References 19 U.S. Copyright Clearance Center Code 0882-1666/91/0012-0079$7.50/0 Country Pub. USA date 1184 ------------------------------------------------------------ Author Lindeberg, T.; Eklundh, J.O.; Comput. Vision & Active Perception Lab., R. Inst. of Technol., Stockholm, Sweden Title Scale detection and region extraction from a scale-space primal sketch Source Proceedings. Third International Conference on Computer Vision (Cat. No.90CH2934-8); Part: Osaka, Japan; Part: 4-7 Dec. 1990; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xv+759; 1990; pp. 416-26 Abstract The authors present: (1) a multi-scale representation of gray- level shape, called a scale-space primal sketch, which makes explicit both features in scale-space and the relations between features at different levels of scales; (2) a theory for extraction of significant image structure from this representation; and (3) applications to edge detection, histogram analysis and junction classification demonstrating how the proposed method can be used for guiding later stage processing. The representation gives a qualitative description of the image structure that allows for detection of stable scales and regions of interest in a solely bottom-up data-driven way. In other words, it generates coarse segmentation cues and can be hence seen as preceding further processing, which can then be properly tuned. The authors argue that once such information is available many other processing tasks can become much simpler. Experiments on real imagery demonstrate that the proposed theory gives perceptually intuitive results Thesaurus computer vision; computerised pattern recognition; computerised picture processing Other Terms scale detection; region extraction; scale-space primal sketch; multi-scale representation; gray-level shape; features; extraction of significant image structure; edge detection; histogram analysis; junction classification; qualitative description; coarse segmentation cues; real imagery ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 4111561 AbstractNos. B9204-6140C-203; C9204-1250-222 ISBN or SBN 0 8186 2057 9 References 30 U.S. Copyright Clearance Center Code CH2934-8/90/0000-0416$01.00 Country Pub. USA date 1182 ------------------------------------------------------------ Author Whitten, G.; Fairchild Weston Syst. Inc., Syosset, NY, USA Title A framework for adaptive scale space tracking solutions to problems in computational vision Source Proceedings. Third International Conference on Computer Vision (Cat. No.90CH2934-8); Part: Osaka, Japan; Part: 4-7 Dec. 1990; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xv+759; 1990; pp. 210-20 Abstract A unified framework is developed for efficiently solving a wide range of computational vision problems by performing adaptive scale space tracking (where a solution at a coarse resolution is tracked to solutions at ever increasing resolution). This approach is motivated by physical smoothness models, deformable sheets, based on thin elastic membranes and plates. The inherent smoothness properties of the deformable sheets act against externally applied, problem specific forces derived from images. The authors also developed the necessary relations for quantitative control of scale based parameters so that the scale space tracking process can be completely automated. Finally, they present solutions to different problems in computational vision using the framework applied to real images Thesaurus computer vision; computerised picture processing Other Terms adaptive scale space tracking; computational vision; coarse resolution; physical smoothness models; deformable sheets; thin elastic membranes; plates; problem specific forces; necessary relations; quantitative control; scale based parameters ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4111527 AbstractNos. B9204-6140C-180; C9204-1250-198 ISBN or SBN 0 8186 2057 9 References 12 U.S. Copyright Clearance Center Code CH2934-8/90/0000-0210$01.00 Country Pub. USA date 1182 ------------------------------------------------------------ Author Deriche, R.; Giraudon, G.; INRIA Sophia-Antipolis, Valbonne, France Title Accurate corner detection: an analytical study Source Proceedings. Third International Conference on Computer Vision (Cat. No.90CH2934-8); Part: Osaka, Japan; Part: 4-7 Dec. 1990; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xv+759; 1990; pp. 66-70 Abstract Consideration is given to a corner model, and its behavior in the scale-space is studied. The authors derive results that clarify the behavior of some well known approaches used to detect corners. In particular, they show that some of the approaches are inadequate for an exact localization of the corner. A novel approach is then proposed in order to correct the displacement effect and detect exactly the corner position. Some promising experimental results obtained on real data are shown Thesaurus computer vision; pattern recognition Other Terms accurate corner detection; computer vision; corner model; scale-space; localization; displacement effect ClassCodes B6140C; C1250 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 4111506 AbstractNos. B9204-6140C-159; C9204-1250-177 ISBN or SBN 0 8186 2057 9 References 13 U.S. Copyright Clearance Center Code CH2934-8/90/0000-0066$01.00 Country Pub. USA date 1182 ------------------------------------------------------------ Author Campos, J.C.; Linney, A.D.; Moss, J.P.; Dept. of Med. Phys., Univ. Coll. London, UK Title The analysis of facial profiles using scale space techniques Source IEE Colloquium on 'Machine Storage and Recognition of Faces' (Digest No.017); Part: London, UK; Part: 24 Jan. 1992; Sponsored by: IEE; London, UK; IEE; 54; 1992; pp. 10/1-3 Abstract Facial description has always been an interesting and intensive topic of research, receiving considerable attention over the past few years in areas such as psychology, bioestereometry, pattern recognition, forensic science, orthodontics and computer vision. Many quantifying techniques and methods of analysis have been proposed. The method developed by the authors has the aim of producing an objective way of identifying landmarks on the human profile leading to a useful segmentation and quantitative description of the contours and features of the face. The main purpose is that of assessing changes in the profile due to surgery or growth, but it could also be used on recognition. The method uses Scale Space techniques, extensively explored in the fields of digital image and signal processing. They make use of filtering the signal across a continuum of scales using Gaussian filters and then tracking the extremal points and their derivatives as they move with scale changes, yielding a useful general purpose qualitative description Thesaurus computerised pattern recognition; computerised picture processing ; medical computing Other Terms digital image processing; facial features; facial profiles; human profile; segmentation; quantitative description; Scale Space techniques; Gaussian filters; qualitative description ClassCodes C5260B; C7330 Article Type Applications; Practical Language English RecordType Conference ControlNo. 4107289 AbstractNos. C9204-5260B-117 References 0 Country Pub. UK date 1197 Title IEE Colloquium on 'Machine Storage and Recognition of Faces' (Digest No.017) Source Part: London, UK; Part: 24 Jan. 1992; Sponsored by: IEE; London, UK; IEE; 54; 1992 Abstract The following topics were dealt with: an architecture for face classification; coding facial images for database retrieval; face recognition algorithm using vector quantization; connectionist model of familiar face recognition; computer-generated cartoons; recognising face features and faces; an extended feature set for automatic face recognition; shape-based description of the facial surface; analysis of facial profiles using scale space techniques; and a technique for generation of facial surface models as aid in orthodontic treatment and orthognathic research Thesaurus computerised pattern recognition; computerised picture processing Other Terms face classification; facial images; database retrieval; face recognition; connectionist model; face features; facial surface ; facial profiles; orthodontic treatment ClassCodes C5260B Article Type Practical Language English RecordType Conference ControlNo. 4107280 AbstractNos. C9204-5260B-111 Country Pub. UK date 1197 ------------------------------------------------------------ Author Lindeberg, T.; Eklundh, J.-O.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title Scale-space primal sketch: construction and experiments Source Image and Vision Computing; Image Vis. Comput. (UK); vol.10, no.1; Jan.-Feb. 1992; pp. 3-18 Abstract Presents a multi-scale representation of grey-level shape, called the scale-space primal sketch, that makes explicit features in scale-space as well as the relations between features at different levels of scale. The representation gives a qualitative description of the image structure that allows for extraction of significant image structure-stable scales and regions of interest- in a solely bottom-up data-driven manner. Hence, it can be seen as preceding further processing, which can then be properly tuned. Experiments on real imagery demonstrate that the proposed theory gives intuitively reasonable results Thesaurus picture processing Other Terms scale detection; segmentation; multi-scale representation; grey-level shape; scale-space primal sketch; real imagery ClassCodes C1250; C5260B Article Type Theoretical / Mathematical Coden IVCODK Language English RecordType Journal ControlNo. 4098695 AbstractNos. C9204-1250-084 ISSN 02628856 References 33 Country Pub. UK date 1197 ------------------------------------------------------------ Author Hong Jeong; Chang-Ik Kim; Woon-Tack Woo; Pohang Inst. of Sci. & Technol., South Korea Title Determining optimal scales for edge detection using regularization Source Proceedings. 1991 IEEE International Conference on Robotics and Automation (Cat. No.91CH2969-4); Part: Sacramento, CA, USA; Part: 9-11 April 1991; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 3 vol. xxxix+2843; 1991; pp. 1596-601 vol.2 Abstract A regularization method for determining the scales of an edge detector adaptively for each part of the image is proposed. The optimal-filter concept of T. Poggio et al. (1984) and the scale space concept of A. Witkin (1983) are extended to an adaptive scale parameter. An energy function defined over continuous scale space is first introduced. Natural constraints for edge detection are incorporated into the energy function. To obtain a set of optimal scales which can minimize the energy function, a parallel relaxation algorithm is introduced. Experiments for synthetic and natural scenes show the advantages of the new algorithm over the edge detectors by B. Marr and E. Hildreth (1980) or J.F. Canny (1986) Thesaurus filtering and prediction theory; optimisation; pattern recognition; picture processing; relaxation theory Other Terms picture processing; pattern recognition; optimal scales; edge detection; regularization method; optimal-filter concept; scale space concept; adaptive scale parameter; energy function; parallel relaxation algorithm ClassCodes B6140C; B0260; C1250; C1260; C1180 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 4097121 AbstractNos. B9204-6140C-038; C9204-1250-056 ISBN or SBN 0 8186 2163 X References 11 U.S. Copyright Clearance Center Code CH2969-4/91/0000-1596$01.00 Country Pub. USA date 1187 ------------------------------------------------------------ Author Nichani, S.; Ranganathan, N.; Dept. of Comput. Sci. & Eng., Univ. of South Florida, Tampa, FL, USA Title SAP: design of a systolic array processor for computation in vision Source Proceedings. 1990 IEEE International Conference on Computer Design: VLSI in Computers and Processors (Cat. No.90CH2909-0); Par t: Cambridge, MA, USA; Part: 17-19 Sept. 1990; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xx+477; 1990; pp. 315-18 Abstract The design and implementation of SAP chip, a systolic array processor for computations in vision, is described. The chip can be used to implement the Gaussian filter, the Laplacian of the Gaussian filter, and scale space generation. The architecture is based on an algorithm that can provide speeds an order of magnitude higher than the speeds of other systems previously proposed. The algorithm utilises the three properties of Gaussian: symmetry, separability, and scaling. The algorithm and the architecture exploit a high degree of pipelining and parallelism in order to obtain high speed, efficiency, and throughput. The architecture is adaptable for masks of any size, and the weights are not restricted to powers of two. The processor was designed using CMOS technology, fabricated, and tested. The chip is fully functional and operates at a rate of 10 MHz Thesaurus CMOS integrated circuits; computer vision; digital signal processing chips; systolic arrays Other Terms vision computation; SAP; systolic array processor; design; implementation; vision; Gaussian filter; Laplacian; scale space generation; symmetry; separability; scaling; pipelining; parallelism; CMOS technology; 10 MHz ClassCodes B1265F; B6140; C5260B Article Type Practical Numerical frequency 1.0E+07 Hz Language English RecordType Conference ControlNo. 4097044 AbstractNos. B9204-1265F-029; C9204-5260B-017 ISBN or SBN 0 8186 2079 X References 14 U.S. Copyright Clearance Center Code CH2909-0/90/0000-0315$01.00 Country Pub. USA date 1179 ------------------------------------------------------------ Author Liu, Z.-Q.; Rangayyan, R.M.; Frank, C.B.; Calgary Univ., Alta., Canada Title Directional analysis of images in scale space Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.13, no.11; Nov. 1991; pp. 1185-92 Abstract The authors propose a computational technique for the directional analysis of piecewise linear patterns in images based on the notion of zero crossings in gradient images. A given image is preprocessed by a sequence of filters that are second derivatives of 2-D Gaussian functions with different scales. This gives a set of zero-crossing maps (the scale space) from which a stability map is generated. Significant linear patterns are detected from measurements on the stability map. Information regarding orientation of the linear patterns in the image and the area covered by the patterns in specific directions is then computed. The performance of the method is illustrated through applications to a simple test image made up of straight bar patterns as well as to scanning electron microscope images of collagen fibrils in rabbit ligaments. The method has significant applications in quantitative analysis of ligament healing and in comparison of treatment methods for ligament injuries Thesaurus biology computing; computer vision; filtering and prediction theory Other Terms scale space images; directional analysis; piecewise linear patterns; zero crossings; gradient images; 2-D Gaussian functions; stability map; orientation; scanning electron microscope images; collagen fibrils; rabbit ligaments; ligament healing; treatment methods; ligament injuries ClassCodes B6140C; C1250; C7330; C1260; C5260B Article Type Bibliography/Literature Suvery; Practical; Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 4093258 AbstractNos. B9204-6140C-019; C9204-1250-025 ISSN 01628828 References 51 U.S. Copyright Clearance Center Code 0162-8828/91/1100-1185$01.00 Country Pub. USA date 1194 ------------------------------------------------------------ Author Li-Dong Cai; Dept. of Artificial Intelligence, Edinburgh Univ., UK Title Spline smoothing: a special case of diffusion smoothing Source Fifth Alvey Vision Conference AVC89. Proceedings of the Fifth Alvey Vision Conference; Part: Reading, UK; Part: 25-28 Sept. 1989; Sheffield, UK; Univ. Sheffield; vi+312; 1989; pp. 273-6 Abstract Diffusion smoothing (DS) implements the smoothing by directly solving a boundary value problem of the diffusion equation delta u/ delta t=b Del /sup 2/u with explicit or implicit numerical schemes, it provides a uniform theoretical base for some other smoothing methods. It has shown that Gaussian smoothing (GS) is equivalent to the initial value problem of DS, and repeated averaging (RA) is a special case of the explicit DS. This paper further proves that spline smoothing (SS) is a special case of the explicit DS with a 'convex corner cling' boundary condition. This result coincides with Poggio's conclusion. However, (1984) the author's proof starts from the diffusion smoothing theory instead of regularisation theory and is given in the mask form; thus it is simpler and more straightforward. Moreover, it makes it possible to explicit the scale-space behaviour of spline smoothing Thesaurus boundary-value problems; computer vision; initial value problems ; partial differential equations; splines (mathematics) Other Terms computer vision; diffusion smoothing; boundary value problem; diffusion equation; numerical schemes; Gaussian smoothing; initial value problem; repeated averaging; spline smoothing; convex corner cling; mask form; scale-space behaviour ClassCodes C4130; C4170; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4081621 AbstractNos. C9203-4130-012 References 9 Country Pub. UK date 1166 ------------------------------------------------------------ Author Mussigmann, U.; Fraunhofer Inst. for Manuf., Eng. & Autom., Stuttgart, Germany Title Homogeneous fractals and their application in texture analysis Source Fractals in the Fundamental and Applied Sciences. Proceedings of the First IFIP Conference; Part: Lisbon, Portugal; Part: 6-8 June 1990; Sponsored by: IFIP; Amsterdam, Netherlands; North-Holland; ix+461; 1991; pp. 269-83 Editor Peitgen, H.-O.; Henriques, J.M.; Penedo, L.F. Abstract The theory of multi-fractals can be used for a complete physical characterization of a fractal set which is inhomogeneous in some sense. The aim of this paper is to prove some basic properties of multi-fractal sets which is done by introducing the local Hausdorff dimension and the local Minkowski dimension. With these definitions of local fractal dimensions one is able to calculate a partition of a fractal set into subsets which are homogeneous with respect to the fractal dimension. For the numerical computation of the fractal dimension the author uses the so- called scale space filtering which has been used in image analysis to describe a discrete signal by its extrema Thesaurus filtering and prediction theory; fractals Other Terms texture analysis; multi-fractals; fractal set; local Hausdorff dimension; local Minkowski dimension; local fractal dimensions; scale space filtering ClassCodes C1160; C1260; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 4077269 AbstractNos. C9203-1160-015 ISBN or SBN 0 444 88757 1 References 16 Country Pub. Netherlands date 1176 ------------------------------------------------------------ Author Yi Lu; Vogt, R.C.; Symbolic Process. Dept., Environ. Res. Inst. of Michigan, Ann Arbor, MI, USA Title Multiscale analysis based on mathematical morphology Source Image Algebra and Morphological Image Processing II; Part: San Diego, CA, USA; Part: 23-24 July 1991; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1568; 1991; pp. 14-25 Abstract The behaviors of objects described by the morphological scale space provide strong knowledge for multiscale analysis. The authors address the two fundamental problems in multiscale analysis: (1) how to select proper scale parameters for various applications, and (2) how to integrate the information filtered at multiscales. They propose two algorithms, Binary Morphological Multiscale Analysis and Gray scale Morphological Multiscale Analysis, for extracting desired regions from binary and gray images Thesaurus computer vision; filtering and prediction theory Other Terms binary images; gray scale images; computer vision; mathematical morphology; morphological scale space; multiscale analysis; scale parameters ClassCodes B6140C; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 4046843 AbstractNos. B9201-6140C-145; C9201-5260B-088 ISSN 0277786X References 11 U.S. Copyright Clearance Center Code 0-8194-0696-1/91/$4.00 Country Pub. USA date 1208 ------------------------------------------------------------ Author Morita, S.; Kawashima, T.; Aoki, Y.; Fac. of Eng., Hokkaido Univ., Sapporo, Japan Title Pattern matching of 2-D shape using hierarchical descriptions Source Systems and Computers in Japan; Syst. Comput. Jpn. (USA); vol.22, no.10; 1991; pp. 40-9 Abstract The paper introduces a system for hierarchical description of two- dimensional shapes on the basis of scale space analysis. The authors propose a group of simple primitives for describing curved line segments; they are suitable for hierarchical analysis. To realize effective matching, nineteen rules are necessary and sufficient to derive a tree. The tree derived from the analysis yields the hierarchical structure of a shape and enables efficient matching of objects in a top-down manner. An algorithm to create a compact database from trees is also shown. The sophisticated database is shown to be useful for recognizing objects by their category. Several examples showed that matching for skewed or occluded shapes can be done by searching for a subtree in the database Thesaurus pattern recognition Other Terms skewed shapes; contour recognition; hierarchical descriptions; two-dimensional shapes; scale space analysis; curved line segments; hierarchical analysis; effective matching; hierarchical structure; occluded shapes ClassCodes C1250; C5260B Article Type Theoretical / Mathematical; Experimental Coden SCJAEP Language English RecordType Journal ControlNo. 4042465 AbstractNos. C9201-1250-099 ISSN 08821666 References 15 U.S. Copyright Clearance Center Code 0882-1666/91/0010-0040$7.50/0 Country Pub. USA date 1184 ------------------------------------------------------------ Author Jepson, A.D.; Fleet, D.J.; Dept. of Comput. Sci., Toronto Univ., Ont., Canada Title Phase singularities in scale-space Source Image and Vision Computing; Image Vis. Comput. (UK); vol.9, no.5; Oct. 1991; pp. 338-43 Abstract The paper concerns the use of phase information from band-pass signals for the measurement of binocular disparity, optic flow and image orientation. Towards this end, one of the important properties of band-pass phase information is its stability with respect to small geometric deformations and contrast changes. However, in particular regions phase can also be very unstable due to the occurrence of phase singularities. The authors discuss the existence of phase singularities, and their relation to the neighbourhoods where phase is unreliable. Moreover, they present a simple method for detecting these regions of instability Thesaurus band-pass filters; computerised picture processing; filtering and prediction theory Other Terms scale-space; band-pass signals; binocular disparity; optic flow ; image orientation; band-pass phase information; stability; geometric deformations; contrast changes; phase singularities; neighbourhoods ClassCodes B6140C; C5260B Article Type Experimental Coden IVCODK Language English RecordType Journal ControlNo. 4033532 AbstractNos. B9201-6140C-051; C9201-5260B-025 ISSN 02628856 References 21 U.S. Copyright Clearance Center Code 0262-8856/91/005338-06$3.00 Country Pub. UK date 1193 ------------------------------------------------------------ Author Lindeberg, T.; Eklundh, J.O.; Comput. Vision & Active Perception Lab., R. Inst. of Technol., Stockholm, Sweden Title On the computation of a scale-space primal sketch Source Journal of Visual Communication and Image Representation; J. Vis. Commun. Image Represent. (USA); vol.2, no.1; March 1991; pp. 55-78 Abstract Scale-space is based on a precise mathematical definition of causality, and the behavior of structure as scale changes can be analytically described. However, the information in the scale- space embedding is only implicit. There is no explicit representation of features or the relations between features at different levels of scale. A theory is presented for constructing such an explicit representation on the basis of formal scale- space theory. The approach is used for gray-level images, but is valid for any bounded function, and can therefore be used to derive properties of, e.g., spatial derivatives. Hence it is useful for studying representations based on intensity discontinuities as well. The representation is obtained in a completely data-driven manner. It gives a description of the image structure that is rather coarse. However, since significant scales and regions are actually determined from the data, the approach can be seen as preceding further processing Thesaurus computerised picture processing Other Terms multiscale representation; blob detection algorithm; computation ; scale-space primal sketch; gray-level images; bounded function; intensity discontinuities; image structure ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical; Experimental Coden JVCRE7 Language English RecordType Journal ControlNo. 4008674 AbstractNos. B91071751; C91067508 ISSN 10473203 References 47 Country Pub. USA date 1186 ------------------------------------------------------------ Author Lindeberg, T.; Eklundh, J.-O.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title Analysis of aerosol images using the scale-space primal sketch Source Machine Vision and Applications; Mach. Vis. Appl. (USA); vol.4, no.3; Summer 1991; pp. 135-44 Abstract The authors outline a method to analyze aerosol images using the scale-space representation. The pictures, which are photographs of an aerosol generated by a fuel injector, contain phenomena that by a human observer are perceived as periodic or oscillatory structures. The presence of these structures is not immediately apparent since the periodicity manifests itself at a coarse level of scale while the dominating objects in the images are small dark blobs, that is, fine scale objects. Experimentally, they illustrate that the scale-space theory provides an objective method to bring out these events. However, in this form the method still relies on a subjective observer in order to extract and verify the existence of the periodic phenomena. Then they extend the analysis by adding a recently developed image analysis concept called the 'scale-space primal sketch'. With this tool, they are able to extract significant structures from a grey-level image automatically without any strong a priori assumptions about either the shape or the scale (size) of the primitives. Experiments demonstrate that the periodic drop clusters they perceived in the image are detected by the algorithm as significant image structures. These results provide objective evidence verifying the existence of oscillatory phenomena Thesaurus aerosols; computerised pattern recognition; computerised picture processing; flow measurement; two-phase flow Other Terms perceptual grouping; blob detection; segmentation; fluid atomization; cluster analysis; aerosol images; scale-space primal sketch; scale-space representation; photographs; fuel injector; periodicity; small dark blobs; fine scale objects; periodic phenomena; image analysis; grey-level image; periodic drop clusters; significant image structures; oscillatory phenomena ClassCodes A4755K; A8270R; A4780; C5260B; C7490; C1250 Article Type Experimental Coden MVAPEO Language English RecordType Journal ControlNo. 4007049 AbstractNos. A91142086; C91070865 ISSN 09328092 References 18 Country Pub. USA date 1184 ------------------------------------------------------------ Author Wada, T.; Yi He Gu; Sato, M.; Res. Lab. of Precision Machinery & Electro., Tokyo Inst. of Technol., Yokohama, Japan Title Scale-space filtering for periodic waveforms Source Systems and Computers in Japan; Syst. Comput. Jpn. (USA); vol.22, no.6; 1991; pp. 45-54 Abstract The paper describes a method to construct scale-space filtering so that the monotonicity of the zero crossings applies also to the periodic waveforms. By analyzing the zero crossings of the waveform, essential differences between ordinary scale-space filtering and periodic scale-space filtering are indicated. An example is presented for the structural analysis of the speech waveform, and the difference between the hierarchical structures derived by those scale-space filters is discussed Thesaurus filtering and prediction theory; speech analysis and processing Other Terms Gaussian filter; periodic waveforms; scale-space filtering; monotonicity; zero crossings; structural analysis; speech waveform; hierarchical structures ClassCodes B6130; C5260 Article Type Theoretical / Mathematical Coden SCJAEP Language English RecordType Journal ControlNo. 4006320 AbstractNos. B91071521; C91070755 ISSN 08821666 References 8 U.S. Copyright Clearance Center Code 0882-1666/91/0006-0045$7.50/0 Country Pub. USA date 1184 ------------------------------------------------------------ Author Freeman, M.O.; Saleh, B.E.A.; Dept. of Electr. & Comput. Eng., Colorado Univ., Boulder, CO, USA Title Centroid scale-space maps Source Journal of the Optical Society of America A (Optics and Image Science); J. Opt. Soc. Am. A, Opt. Image Sci. (USA); vol.8, no.9; Sept. 1991; pp. 1474-87 Abstract An image representation based on mapping image centroids as a function of spatial and scale variables is introduced. This new representation has the advantage that the contours within this scale-space coordinate system consist primarily of curves (i.e. one-dimensional forms) even when the input function is an image with two (or more) spatial coordinates. This is contrasted with the more well-known scale-space approaches based on tracking edge locations through scale when the scale-space contours become curved surfaces for two-dimensional input images. The one- dimensional form of the centroid scale-space contours simplifies the process of extracting useful information from the scale-space maps. Centroid scale-space maps for one- and two-dimensional input functions are examined, and a preliminary set of properties relating the inputs to the maps is presented. The author demonstrates the applications of finding the locations and size distributions of objects in an image and of determining the medial-axis transform of an image Thesaurus picture processing Other Terms centroid scale space maps; image representation; image centroids ; scale variables; scale-space coordinate system; curves; one- dimensional forms; spatial coordinates; centroid scale-space contours; two-dimensional input functions; locations; size distributions; objects; medial-axis transform ClassCodes A4230V Article Type Theoretical / Mathematical; Experimental Coden JOAOD6 Language English RecordType Journal ControlNo. 4003735 AbstractNos. A91141365 ISSN 07403232 References 17 U.S. Copyright Clearance Center Code 0740-3232/91/091474-14$05.00 Country Pub. USA date 1192 ------------------------------------------------------------ Author ter Haar Romeny, B.M.; Florack, L.M.J.; Koenderink, J.J.; Viergever, M.A.; Utrecht Univ. Hospital, Netherlands Title Scale space: its natural operators and differential invariants Source Information Processing in Medical Imaging. 12th International Conference, IPMI '91 Proceedings; Part: Wye, UK; Part: 7-12 July 1991; Berlin, Germany; Springer-Verlag; xi+512; 1991; pp. 239-55 Editor Colchester, A.C.F.; Hawkes, D.J. Abstract The authors discuss the fundamental concept of scaling as well as some natural constraints of a front-end visual system and show that a complete hierarchical set of scaled differential operators follows from these considerations. The lowest order kernel is the isotropic Gaussian. The higher order kernels are the scaled Gaussian directional derivatives, which form the natural, scaled differential operators. With this set they can study local image geometry to any desired order. To this end they introduce the concept of a local jet of order N, J/sup N/(L(P)), also called N- jet (Poston and Steward 1978), defined as the equivalence class of functions L which share the same N-truncated Taylor expansion at a given point P Thesaurus invariance; picture processing Other Terms scaling; front-end visual system; scaled differential operators; lowest order kernel; isotropic Gaussian; scaled Gaussian directional derivatives; local image geometry; local jet; N-jet ; equivalence class; N-truncated Taylor expansion ClassCodes B6140C; C1250; C5260B Language English RecordType Conference ControlNo. 4000432 AbstractNos. B91071828; C91062245 ISBN or SBN 3 540 54246 9 References 29 Country Pub. Germany date 1190 ------------------------------------------------------------ Author Amirfathi, M.M.; Morris, S.; O'Rorke, P.; Bond, W.E.; St. Clair, D.C.; Douglas Aircraft Co., Long Beach, CA, USA Title Pattern recognition for nondestructive evaluation Source 1991 IEEE Aerospace Applications Conference Digest (Cat. No. 91TH0361-6); Part: Crested Butte, CO, USA; Part: 3-8 Feb. 1991; Sponsored by: IEEE; New York, NY, USA; IEEE; 242; 1991; pp. 6/1-11 Abstract The issues involved in automating nondestructive evaluation (NDE) techniques are outlined. Attention is given to research focused on the application of machine learning techniques to the construction and maintenance of knowledge-based systems which are capable of evaluating the readings from nondestructive tests that have been performed on aircraft components. Preliminary results obtained from this research are described. In particular, the authors discuss the application of a symbolic machine learning algorithm, ID3, to the NDE problem. ID3 has been used by Douglas Aircraft to classify defects in sets of standard NDE reference blocks. Based on the preliminary results, a need for an improved method of distinguishing features in the test waveforms is identified. The authors also outline a feature extraction approach from pattern recognition, called scale-space filtering, which can be used to preprocess data for input into a classification algorithm such as ID3 Thesaurus automatic testing; computerised pattern recognition; knowledge based systems; learning systems; mechanical engineering computing; nondestructive testing Other Terms data preprocessing; A-scan; composite materials; artificial intelligence; NDT; nondestructive evaluation; knowledge-based systems; nondestructive tests; aircraft components; symbolic machine learning algorithm; Douglas Aircraft; test waveforms; feature extraction; pattern recognition; scale-space filtering; classification algorithm ClassCodes B0590; B7210B; B7620; C7410H; C5260B; C6170; C7440 Article Type Practical Language English RecordType Conference ControlNo. 4000198 AbstractNos. B91068015; C91066202 ISBN or SBN 0 87942 686 1 References 11 U.S. Copyright Clearance Center Code TH0361-6/91/0000-0001$01.00 Country Pub. USA date 1185 ------------------------------------------------------------ Author Saint-Marc, P.; Chen, J.-S.; Medioni, G.; Inst. for Robotics & Intelligent Syst., Univ. of Southern California, Los Angeles, CA, USA Title Adaptive smoothing: a general tool for early vision Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.13, no.6; June 1991; pp. 514-29 Abstract A method to smooth a signal while preserving discontinuities is presented. This is achieved by repeatedly convolving the signal with a very small averaging mask weighted by a measure of the signal continuity at each point. Edge detection can be performed after a few iterations, and features extracted from the smoothed signal are correctly localized (hence, no tracking is needed). This last property allows the derivation of a scale-space representation of a signal using the adaptive smoothing parameter k as the scale dimension. The relation of this process to anisotropic diffusion is shown. A scheme to preserve higher-order discontinuities and results on range images is proposed. Different implementations of adaptive smoothing are presented, first on a serial machine, for which a multigrid algorithm is proposed to speed up the smoothing effect, then on a single instruction multiple data (SIMD) parallel machine such as the Connection Machine. Various applications of adaptive smoothing such as edge detection, range image feature extraction, corner detection, and stereo matching are discussed Thesaurus adaptive filters; computer vision; computerised pattern recognition; computerised picture processing; filtering and prediction theory Other Terms machine vision; computer vision; pattern recognition; adaptive filtering; scale-space representation; adaptive smoothing; anisotropic diffusion; SIMD; parallel machine; Connection Machine; edge detection; feature extraction; corner detection; stereo matching ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3991725 AbstractNos. B91071716; C91062173 ISSN 01628828 References 39 U.S. Copyright Clearance Center Code 0162-8828/91/0600-0514$01.00 Country Pub. USA date 1189 ------------------------------------------------------------ Author Ueda, N.; Suzuki, S.; Human Interface Labs., NTT, Yokosuka, Japan Title A matching algorithm of deformed planar curves using multiscale convex/concave structures Source Systems and Computers in Japan; Syst. Comput. Jpn. (USA); vol.22, no.5; 1991; pp. 94-104 Abstract The paper proposes a new multiscale segment matching method which is applicable to heavily deformed planar shapes. First, multiscale representations are obtained using curvature scale space filtering. Then inflection point correspondence is developed between consecutive smoothed shapes. The representation in this paper, unlike the well-known curvature scale space image description, ensures that it always satisfies the consistency of hierarchical segment replacement. Moreover, it requires less processing time and memory allocation. Finally, optimum scale segments are matched by a new multiscale segment matching method proposed herein. In this method, the matching problem is formulated as a minimization problem of the total amount of segment dissimilarity. The minimization problem is solved effectively using dynamic programming. The proposed matching method makes it possible to obtain intuitively relevant correspondences even if the shapes have some local heavy deformation Thesaurus computerised pattern recognition; filtering and prediction theory Other Terms matching algorithm; deformed planar curves; multiscale convex/concave structures; multiscale segment matching method; heavily deformed planar shapes; multiscale representations; curvature scale space filtering; inflection point correspondence; smoothed shapes; hierarchical segment replacement; processing time; memory allocation; minimization problem; segment dissimilarity; dynamic programming; intuitively relevant correspondences; local heavy deformation ClassCodes C5260B; C1250; C1260 Article Type Theoretical / Mathematical Coden SCJAEP Language English RecordType Journal ControlNo. 3982095 AbstractNos. C91064000 ISSN 08821666 References 4 U.S. Copyright Clearance Center Code 0882-1666/91/0005-0094$7.50/0 Country Pub. USA date 1184 ------------------------------------------------------------ Author Liu, Z.-Q.; Rangayyan, R.M.; Frank, C.B.; Novatel Commun. Ltd., Calgary, Alta., Canada Title Statistical analysis of collagen alignment in ligaments by scale- space analysis Source IEEE Transactions on Biomedical Engineering; IEEE Trans. Biomed. Eng. (USA); vol.38, no.6; June 1991; pp. 580-8 Abstract The authors propose a computational technique for statistical analysis of collagen alignment in ligament images using the scale- space approach. In this method, a ligament image is preprocessed by a sequence of filters which are second derivatives of two- dimensional Gaussian functions with different scales. This gives a set of zero-crossing maps (the scale space) from which a stability map is generated. Significant linear patterns are captured by analyzing the stability map. The directional information in terms of orientation distributions of the collagen fibrils in the image and the area covered by the fibrils in specific directions is extracted for statistical analysis. Examples illustrating the performance of this method with scanning electron microscope images of the collagen fibrils in healing rabbit medial collateral ligaments are presented Thesaurus picture processing; proteins; statistical analysis Other Terms 2D Gaussian functions; image preprocessing; filters sequence; collagen alignment; scale-space analysis; computational technique; zero-crossing maps; linear patterns; stability map; collagen fibrils; scanning electron microscope images; healing rabbit medial collateral ligaments ClassCodes A8710 Article Type Theoretical / Mathematical Coden IEBEAX Language English RecordType Journal ControlNo. 3971104 AbstractNos. A91123964 ISSN 00189294 References 32 U.S. Copyright Clearance Center Code 0018-9294/91/0600-0580$01.00 Country Pub. USA date 1189 ------------------------------------------------------------ Author Cumani, A.; Grattoni, P.; Guiducci, A.; Istituto Elettrotecnico Nazionale Galileo Ferraris, Torino, Italy Title An edge-based description of color images Source CVGIP: Graphical Models and Image Processing; CVGIP, Graph. Models Image Process. (USA); vol.53, no.4; July 1991; pp. 313-23 Abstract This paper presents a method of describing multispectral images, for computer vision applications, in terms of contour elements. Contours are detected, at different scales of resolution, as the zero crossings of a second-order differential operator that represents an extension of the second directional derivative to the multispectral case. A fine-to-coarse analysis of contour behavior in scale space is then used to compute the attributes needed for the description of the image. Subsequent contour segmentation, based on both geometric and photometric features, allows for a further increase in description compactness without significant losses of information. In order to assess the faithfulness of the description, it is shown that an approximate reconstruction of the original image can be obtained from the coded contours. The method has been tested on several real-world colour images. Two examples, in which images are described and reconstructed at different degrees of compactness, allowing for an objective and subjective evaluation of the performance of the method, are presented Thesaurus computer vision; computerised picture processing Other Terms edge-based description; color images; multispectral images; computer vision; contour elements; zero crossings; second- order differential operator; fine-to-coarse analysis; coded contours ClassCodes C1250; C5260B Article Type Practical Coden CGMPE5 Language English RecordType Journal ControlNo. 3964020 AbstractNos. C91055741 ISSN 10499652 References 22 Country Pub. USA date 1190 ------------------------------------------------------------ Author Wada, T.; Sato, M.; Res. Lab. of Precision Machinery & Electron., Tokyo Inst. of Technol., Japan Title Scale-space tree and its hierarchy Source Proceedings. 10th International Conference on Pattern Recognition (Cat. No.90CH2898-5); Part: Atlantic City, NJ, USA; Part: 16-21 June 1990; Sponsored by: Int. Assoc. Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 2 vol. (xxxi+xxv+1676); 1990; pp. 103-8 vol.2 Abstract Scale-space filtering is a multiresolutional filtering method which expands a waveform to the set of waveforms in multiscale measurement. The intervals bounded by zero-crossings of expanded waveforms have a hierarchical structure, which can be represented by an interval tree. Another hierarchical description of the waveform, scale-space tree, is introduced as a natural description of the hierarchy. The difference hierarchy between the interval tree and the scale-space tree is discussed Thesaurus filtering and prediction theory; picture processing Other Terms scale-space filtering; multiresolutional filtering; waveforms; multiscale measurement; zero-crossings; hierarchical structure; interval tree; scale-space tree ClassCodes B6140C; C1260; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3957030 AbstractNos. B91058384; C91051183 ISBN or SBN 0 8186 2062 5 References 8 U.S. Copyright Clearance Center Code CH2898-5/90/0000-0103$01.00 Country Pub. USA date 1176 ------------------------------------------------------------ Author Woo Young Choi; Jong Soo Choi; San Uk Lee; Rae Hong Park; Dept. of Electr. Eng., Sogang Univ., Seoul, South Korea Title Stereo matching using finger print on the scale space Source Journal of the Korean Institute of Telematics and Electronics; J. Korean Inst. Telemat. Electron. (South Korea); vol.28B, no.2; Feb. 1991; pp. 53-60 Abstract Proposes a stereo correspondence matching algorithm using the finger print (characteristic pattern) of zero-crossing points on the scale-space as the robust feature. In the authors' approach they extract the finger print corresponding to the authentic zero- crossing edge along the scale value due to the maximum principle. Since most of the objects cannot be matched by only their local feature values, they used the relaxation matching method using not only the feature values but also the available contextual information. In simulation, they applied the proposed algorithm to synthetic and natural images and obtained good matching results Thesaurus computerised picture processing Other Terms pattern of zero-crossing points; stereo image matching; synthetic images; stereo correspondence matching algorithm; scale-space; zero-crossing edge; relaxation matching method; contextual information; natural images; matching results ClassCodes B6140C; C5260B; C7410F Article Type Theoretical / Mathematical; Experimental Coden CKNOEZ Language English RecordType Journal ControlNo. 3953479 AbstractNos. B91058192; C91052608 References 12 Country Pub. South Korea date 1185 ------------------------------------------------------------ Author Raman, S.V.; Sarkar, S.; Boyer, K.L.; Dept. of Electr. Eng., Ohio State Univ., Columbus, OH, USA Title Tissue boundary refinement in magnetic resonance images using contour-based scale space matching Source IEEE Transactions on Medical Imaging; IEEE Trans. Med. Imaging (USA); vol.10, no.2; June 1991; pp. 109-21 Abstract An algorithm for computationally focusing the tissue boundaries detected from magnetic resonance images is presented. The proposed approach is a novel, whole-contour-based technique for tracing edges selected at a coarse scale into successively finer scales to recover the needed precision. The tracing algorithm builds consensus through a fast pixel voting scheme. Also presented is a rigorous method for determining the appropriate itinerary when traversing scale space, beginning from the premise of a maximum pixel migration per unit change in scale parameter. This leads to an efficient method of processing images so as to maximize accuracy and minimize noise. Although the LoG (Laplacian of Gaussian) is used for many of the experiments, results using a novel edge detector which is mathematically superior to and faster to compute than the LoG and for which fewer steps are required to traverse the same effective span in scale space are presented. Experimental results on real data are presented, and other potential applications are discussed Thesaurus biomedical NMR; picture processing Other Terms tissue boundary refinement; edge tracing; fine scale; magnetic resonance images; contour-based scale space matching; algorithm; whole-contour-based technique; coarse scale; fast pixel voting scheme; itinerary; scale parameter; Laplacian of Gaussian ClassCodes A8760G; A8770E; A8740 Article Type Theoretical / Mathematical Coden ITMID4 Language English RecordType Journal ControlNo. 3947079 AbstractNos. A91112501 ISSN 02780062 References 23 U.S. Copyright Clearance Center Code 0278-0062/91/0600-0109$01.00 Country Pub. USA date 1189 ------------------------------------------------------------ Author Wilson, R.; Clippingdale, S.; Bhalerao, A.H.; Dept. of Comput. Sci., Warwick Univ., Coventry, UK Title Robust estimation of local orientations in images using a multiresolution approach Source Visual Communications and Image Processing '90; Part: Lausanne, Switzerland; Part: 1-4 Oct. 1990; Sponsored by: SPIE; Swiss Federal Inst. Technol; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1360, pt.3; 1990; pp. 1393-1403 Abstract The problem of estimating feature orientation from noisy image data is addressed. It is shown that by appropriate choice of representation of orientation, it is possible to employ simple linear smoothing methods to reduce estimation noise. A combination of scale-space recursive filtering and iterative estimation gives significant improvements in estimated orientations at low computational cost. Applications to enhancement and are presented Thesaurus estimation theory; filtering and prediction theory; iterative methods; picture processing Other Terms multiresolution approach; feature orientation; noisy image data; linear smoothing; scale-space recursive filtering; iterative estimation; computational cost; enhancement ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3946751 AbstractNos. B91058279; C91050880 ISSN 0277786X References 18 Country Pub. USA date 1180 ------------------------------------------------------------ Author Ren, Z.; Ameling, W.; Jensch, P.; Rogowski-Inst. fur Elektrotech., Tech. Hochschule Aachen, Germany Title An attributed tree data structure for representing the descriptions of object contours in images Source Visual Communications and Image Processing '90; Part: Lausanne, Switzerland; Part: 1-4 Oct. 1990; Sponsored by: SPIE; Swiss Federal Inst. Technol; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1360, pt.2; 1990; pp. 956-69 Abstract The shapes of medical objects, e.g. human bones or skeletons, conveys information about anatomic structures and pathological states. The sectional contours of objects observable through computer tomograms provide descriptive clues for analyzing shapes. Sectional contours of medical objects are highly curved and complex and therefore should be described in a higher order parameter space. These considerations have guided the authors to examine the scale-space transformation, and to develop methods for expressing, describing, and extracting structures in the scale-space images of the contours. This paper introduces a tree- based data structure for representing descriptions of structures in the scale-space, and further demonstrates its applications to shape analysis Thesaurus computerised tomography; data structures; medical diagnostic computing; trees (mathematics) Other Terms descriptions of object contours; medical objects; sectional contours; computer tomograms; scale-space transformation; tree- based data structure; shape analysis ClassCodes A8760J; A8710; B6140C; C7330; C5260B; C6120 Article Type Applications; Practical Coden PSISDG Language English RecordType Conference ControlNo. 3946721 AbstractNos. A91104852; B91058253; C91054342 ISSN 0277786X References 30 Country Pub. USA date 1180 ------------------------------------------------------------ Author Wilson, R.; Todd, M.; Calway, A.D.; Dept. of Comput. Sci., Warwick Univ., Coventry, UK Title Generalized quad-trees: a unified approach to multiresolution image analysis and coding Source Visual Communications and Image Processing '90; Part: Lausanne, Switzerland; Part: 1-4 Oct. 1990; Sponsored by: SPIE; Swiss Federal Inst. Technol; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1360, pt.1; 1990; pp. 619-26 Abstract This paper is an attempt to bring a number of current ideas in multiresolution image processing into a single framework. The unification is achieved by a two-level image model, comprising a quadtree containing parameters which control the evolution of the image as a sequence of successive refinements through scale-space. Examples of the method's application to image coding and analysis illustrate the principles and show its usefulness Thesaurus encoding; hierarchical systems; picture processing Other Terms multiresolution image analysis; image processing; two-level image model; quadtree; sequence of successive refinements; image coding ClassCodes B6140C; B6120B; C1250 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3946707 AbstractNos. B91058239; C91050842 ISSN 0277786X References 18 Country Pub. USA date 1180 ------------------------------------------------------------ Author Rattarangsi, A.; Chin, R.T.; Dept. of Electr. & Comput. Eng., Wisconsin Univ., Madison, WI, USA Title Scale-based detection of corners of planar curves Source Proceedings. 10th International Conference on Pattern Recognition (Cat. No.90CH2898-5); Part: Atlantic City, NJ, USA; Part: 16-21 June 1990; Sponsored by: Int. Assoc. Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 2 vol. (xxxi+xxv+1676); 1990; pp. 923-30 vol.1 Abstract A technique for detecting and localizing corners of planar curves is proposed. The technique is based on Gaussian scale space, which consists of the maxima of absolute curvature of the boundary function presented at all scales. The scale space of isolated simple and double corners is analyzed to investigate the behavior of scale space due to smoothing and interactions between two adjacent corners. The scale space is transformed into a tree which provides simple but concise representation of corners at multiple scales. A multiple-scale corner detection scheme is developed using a coarse-to-fine tree parsing technique. The parsing scheme is based on a stability criterion which states that the presence of a corner must concur with a curvature maximum observable at a majority of scales. Experimental results show that the scale-space corner detector is reliable for objects with multiple-size features and noisy boundaries and that it compares favorably with other corner detectors tested Thesaurus grammars; pattern recognition; picture processing; trees (mathematics) Other Terms planar curve corner detection; Gaussian scale space; absolute curvature; boundary function; scale space; tree parsing; stability criterion; scale-space corner detector ClassCodes B6140C; B0250; C1250; C1160; C4210 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3944495 AbstractNos. B91051673; C91051046 ISBN or SBN 0 8186 2062 5 References 20 U.S. Copyright Clearance Center Code CH2898-5/90/0000-0923$01.00 Country Pub. USA date 1176 ------------------------------------------------------------ Author Wu, Y.; Maitre, H.; Telecom, Paris, France Title Registration of a SPOT image and a SAR image using multiresolution representation of a coastline Source Proceedings. 10th International Conference on Pattern Recognition (Cat. No.90CH2898-5); Part: Atlantic City, NJ, USA; Part: 16-21 June 1990; Sponsored by: Int. Assoc. Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 2 vol. (xxxi+xxv+1676); 1990; pp. 913-17 vol.1 Abstract A novel method is presented for the registration of a SPOT satellite image against a Seasat synthetic aperture radar (SAR) image. An edge following technique is proposed. It is adapted to very noisy images such as SAR images. It is shown that, by the use of a multiresolution method, a directional graph can be constructed which gives a clearer definition of the equivalence between edge following and optimal path search on a graph. A hypothesis-testing approach, also based on a multiresolution method, is proposed to solve the registration problem. The hypotheses are drawn from scale space representations in order to maintain a low complexity. These hypotheses are then tested by a chamfer distance map. By the combination of scale space representation and the chamfer matching method, a very robust and fast registration results Thesaurus computerised picture processing; geophysical techniques; graph theory; oceanographic techniques; remote sensing; remote sensing by radar Other Terms remote sensing; computerised picture processing; SPOT image; SAR image; multiresolution representation; coastline; edge following technique; directional graph; optimal path search; scale space representations; chamfer distance map ClassCodes A9385; A9210; A9190; B7710; B7730; B6140C; B0250; C7340; C5260B; C1160 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3944493 AbstractNos. A91100025; B91053354; C91054395 ISBN or SBN 0 8186 2062 5 References 21 U.S. Copyright Clearance Center Code CH2898-5/90/0000-0913$01.00 Country Pub. USA date 1176 ------------------------------------------------------------ Author Ueda, N.; Suzuki, S.; NTT Corp., Yokosuka, Japan Title Automatic shape model acquisition using multiscale segment matching Source Proceedings. 10th International Conference on Pattern Recognition (Cat. No.90CH2898-5); Part: Atlantic City, NJ, USA; Part: 16-21 June 1990; Sponsored by: Int. Assoc. Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 2 vol. (xxxi+xxv+1676); 1990; pp. 897-902 vol.1 Abstract A novel method for acquiring a shape model from shape samples of the same class is proposed. A critical point is that the method requires no prior knowledge of the class. Multiscale representations are first obtained using curvature scale space filtering to gain inflection point correspondence between consecutive smoothed shapes. The multiscale samples are then matched to extract the convex/concave structure common to the class. The matching is invariant under translation, rotation, and size change. Finally, generalized samples composing a model are generated by smoothly connecting the matched convex and concave segments. Experimental results show that the resulting model is useful for shape recognition Thesaurus filtering and prediction theory; pattern recognition; picture processing Other Terms multiscale representation; picture processing; pattern recognition; shape model acquisition; multiscale segment matching; curvature scale space filtering; convex/concave structure; shape recognition ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3944490 AbstractNos. B91051670; C91051042 ISBN or SBN 0 8186 2062 5 References 11 U.S. Copyright Clearance Center Code CH2898-5/90/0000-0897$01.00 Country Pub. USA date 1176 ------------------------------------------------------------ Author Allmen, M.; Dyer, C.R.; Dept. of Comput. Sci., Wisconsin Univ., Madison, WI, USA Title Cyclic motion detection using spatiotemporal surfaces and curves Source Proceedings. 10th International Conference on Pattern Recognition (Cat. No.90CH2898-5); Part: Atlantic City, NJ, USA; Part: 16-21 June 1990; Sponsored by: Int. Assoc. Pattern Recognition; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 2 vol. (xxxi+xxv+1676); 1990; pp. 365-70 vol.1 Abstract Cyclic motion is formally defined as repeating curvature values along a path of motion. A procedure is presented for cyclic motion detection using spatiotemporal (ST) surfaces and ST curves. The projected movement of an object generates ST surfaces. ST curves are detected on the ST surfaces, providing an accurate, compact, qualitative description of the ST surfaces. Curvature scale-space of the ST curves is then used to detect intervals of repeating curvature values. The successful detection of cyclic motion in two data sets is presented Thesaurus pattern recognition; visual perception Other Terms spatiotemporal curves; curvature scale-space; spatiotemporal surfaces; cyclic motion detection; repeating curvature values ClassCodes A8732S; B6140C; C1250; C1290L Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3944398 AbstractNos. A91104717; B91051616; C91050975 ISBN or SBN 0 8186 2062 5 References 19 U.S. Copyright Clearance Center Code CH2898-5/90/0000-0365$01.00 Country Pub. USA date 1176 ------------------------------------------------------------ Author Ueda, N.; Suzuki, S.; NTT Human Interface Lab., Tokyo, Japan Title Multiscale convex/concave structure matching -MC matching method Source NTT R & D; NTT R & D (Japan); vol.40, no.3; 1991; pp. 399-406 Abstract Proposes a new multiscale segment matching method that is applicable to heavily deformed planar shapes. First, multiscale representations are obtained using curvature scale space filtering, which requires less processing time and memory allocation than the well-known curvature scale space image description. Then, inflection point correspondence is developed between consecutive smoothed shapes. Finally, optimum scale segment correspondences are determined from all possible combinations of the two multiscale representations. The procedure is performed effectively using dynamic programming. The proposed matching method makes it possible to obtain intuitively relevant correspondences even if the shapes have some local heavy deformation Thesaurus computerised pattern recognition; dynamic programming Other Terms multiscale convex structure matching; multiscale concave structure matching; multiscale segment matching; deformed planar shapes; curvature scale space filtering; inflection point ; optimum scale segment correspondences; dynamic programming ClassCodes B6140C; B0260; C5260B; C1180 Article Type Practical Coden NTTDEC Language Japanese RecordType Journal ControlNo. 3942045 AbstractNos. B91051494; C91052611 ISSN 09152326 References 7 Country Pub. Japan date 1184 ------------------------------------------------------------ Author Zhi-Qiang Liu; Dept. of Adv. Digital Process., Novatel Commun. Ltd., Calgary, Alta., Canada Title Scale space approach to directional analysis of images Source Applied Optics; Appl. Opt. (USA); vol.30, no.11; 10 April 1991; pp. 1369-73 Abstract A new technique for directional analysis of linear patterns in images is proposed based on the notion of scale space. A given image is preprocessed by a sequence of filters which are second derivatives of 2-D Gaussian functions with different scales. This gives a set of zero crossing maps (the scale space) from which a stability map is generated. Significant linear patterns are detected from measurements on the stability map. Information regarding orientation of the linear patterns in the image and the area covered by the patterns in specific directions is then computed. The performance of the method is illustrated through applications to synthetic patterns and to scanning electron microscope images of collagen fibrils in rabbit ligaments Thesaurus biological techniques and instruments; computerised pattern recognition; computerised picture processing; scanning electron microscope examination of materials Other Terms filter sequence; 2D Gaussian function second derivatives; computer vision; directional analysis; images; linear patterns; scale space; zero crossing maps; stability map; linear patterns; synthetic patterns; scanning electron microscope images; collagen fibrils; rabbit ligaments ClassCodes A0650D; A4230V; A0780; A8780; A4230S; B6140C; C1250 Article Type New Development; Practical; Theoretical / Mathematical; Experimental Coden APOPAI Language English RecordType Journal ControlNo. 3928839 AbstractNos. A91089235; B91051452; C91044927 ISSN 00036935 References 23 U.S. Copyright Clearance Center Code 0003-6935/91/111369-05$05.00/0 Country Pub. USA date 1187 ------------------------------------------------------------ Author Vincken, K.L.; de Graaf, C.N.; Koster, A.S.E.; Viergever, M.A.; Appelman, F.J.R.; Timmens, G.R.; Utrecht Univ., Netherlands Title Multiresolution segmentation of 3D images by the hyperstack Source Proceedings of the First Conference on Visualization in Biomedical Computing (Cat. No.90TH0311-1); Part: Atlanta, GA, USA ; Part: 22-25 May 1990; Sponsored by: IEEE; Georgia Inst. Technol.; NSF; Emory Univ. School of Med.; et al; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xii+520; 1990; pp. 115-22 Abstract An improvement in the design of the hyperstack, a three- dimensional image segmentation tool, is described. It is based on a multiresolution approach which is mathematically supported by the diffusion equation. The blurring strategy, used to build the scale space, is outlined, including some difficulties that occur in view of the transition from 2-D to 3-D. The existing prototype, a hyperstack grounded on isointensity following, is extended by two novel ideas: weighted linking and stand-alone parents. The result of a segmented 3D SPECT image (of a liver) is shown. Theoretical considerations concerning the addition of feature information to guide the segmentation process are briefly mentioned. A flexible way to obtain several output images from one single hyperstack is outlined and the reduction of the sampling rate by means of interpolation, which will decrease the total amount of processing time, is investigated Thesaurus computerised pattern recognition; computerised picture processing ; data structures; medical computing Other Terms data structure; 3D images; hyperstack; three-dimensional image segmentation tool; multiresolution; diffusion equation; blurring strategy; scale space; isointensity following; weighted linking; stand-alone parents; segmented 3D SPECT image; liver; feature information; segmentation; sampling rate; interpolation; processing time ClassCodes C5260B; C7330 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3926454 AbstractNos. C91047206 ISBN or SBN 0 8186 2039 0 References 10 U.S. Copyright Clearance Center Code TH0311-1/90/0000-0115$01.00 Country Pub. USA date 1175 ------------------------------------------------------------ Author Cullip, T.J.; Fredericksen, R.E.; Gauch, J.M.; Pizer, S.M.; North Carolina Univ., Chapel Hill, NC, USA Title Algorithms for 2D and 3D image description based on the IAS Source Proceedings of the First Conference on Visualization in Biomedical Computing (Cat. No.90TH0311-1); Part: Atlanta, GA, USA ; Part: 22-25 May 1990; Sponsored by: IEEE; Georgia Inst. Technol.; NSF; Emory Univ. School of Med.; et al; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; xii+520; 1990; pp. 102-7 Abstract The algorithms used in computing the intensity axis of symmetry (IAS) for 2D and 3D medical images are described. The basic 2D algorithms, along with the algorithms needed to incorporate scale space, are described. A brief discussion of the extensions needed to work with 3D images is given. The basic approach is to treat the image as a deformable intensity surface which is contracted onto the IAS. The primitive regions of the segmentation are identified by the branches in the resulting tree-like structure. A hierarchy is produced by following the simplification of the branching through scale space. For comparison papers see ibid., p. 94-101 and ibid., p.108-14 Thesaurus computerised picture processing; medical computing Other Terms CT image; chest; MRI image; brain; ridge structures; branching structures; IAS segmentation; image segmentation; 2D and 3D image description; IAS; intensity axis of symmetry; medical images; deformable intensity surface; primitive regions; segmentation; branching; scale space ClassCodes C5260B; C7330 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3926452 AbstractNos. C91047204 ISBN or SBN 0 8186 2039 0 References 3 U.S. Copyright Clearance Center Code TH0311-1/90/0000-0102$01.00 Country Pub. USA date 1175 ------------------------------------------------------------ Author Bengtsson, A.; Eklundh, J.-O.; Comput. Vision & Active Perception Lab., R. Inst. of Technol., Stockholm, Sweden Title Shape representation by multiscale contour approximation Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.13, no.1; Jan. 1991; pp. 85-93 Abstract An approach is presented for deriving qualitative descriptions of contours containing structures at different (unknown) scales. The descriptions are in terms of straight arcs, curved arcs with sign of curvature, corners, and points delimiting the arcs: inflexion points and transitions from straight to curved. Furthermore, the tangents at these points are derived. The approach is based on the construction of a hierarchic family of polygons, having the scale-space property of causality; structure can only disappear as scale goes from fine to coarse. Using the principle that structures that are stable over scale represent significant properties, the features of the descriptive representations are then derived Thesaurus computer vision; computerised pattern recognition; computerised picture processing Other Terms shape representation; scale stability; multiscale contour approximation; qualitative descriptions; straight arcs; curved arcs; corners; points delimiting the arcs; inflexion points; polygons; scale-space property of causality ClassCodes C5260; C1250 Article Type Theoretical / Mathematical; Experimental Coden ITPIDJ Language English RecordType Journal ControlNo. 3880519 AbstractNos. C91035561 ISSN 01628828 References 27 U.S. Copyright Clearance Center Code 0162-8828/91/0100-0085$01.00 Country Pub. USA date 1184 ------------------------------------------------------------ Author Estola, K.-P.; Nokia Mobile Phones, Salo, Finland Title Interpolated Gaussian scale-space filters Source ICASSP 90. 1990 International Conference on Acoustics, Speech and Signal Processing (Cat. No.90CH2847-2); Part: Albuquerque, NM, USA; Part: 3-6 April 1990; Sponsored by: IEEE; New York, NY, USA; IEEE; 5 vol. 2970; 1990; pp. 2073-6 vol.4 Abstract Computationally efficient interpolated Gaussian scale-space filters are introduced. The proposed scale-space filters incorporate multirate signal-processing methods and are realized using multistage filter structures. The scale-space filters can also be used in decimating the data, if desired. The proposed filtering methods require dramatically less computation than conventional Gaussian scale-space filtering, especially where the scale changes with an integer factor. The filters are extremely efficient when the change in scale is an integer power of an integer constant Thesaurus filtering and prediction theory; interpolation; signal processing Other Terms interpolated Gaussian scale-space filters; multirate signal- processing methods; multistage filter structures ClassCodes B6140; B0290F Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3875736 AbstractNos. B91033196 References 4 U.S. Copyright Clearance Center Code CH2847-2/90/0000-2073$01.00 Country Pub. USA date 1174 ------------------------------------------------------------ Author Zuerndorfer, B.; Wakefield, G.H.; England, A.W.; Dept. of Electr. Eng. & Comput. Sci., Michigan Univ., Ann Arbor, MI, USA Title Recovery of fine resolution information in multispectral processing Source ICASSP 90. 1990 International Conference on Acoustics, Speech and Signal Processing (Cat. No.90CH2847-2); Part: Albuquerque, NM, USA; Part: 3-6 April 1990; Sponsored by: IEEE; New York, NY, USA; IEEE; 5 vol. 2970; 1990; pp. 2033-6 vol.4 Abstract Multiple-sensor processing is considered, and a unified method for representing multiple-sensor data is developed. When resolution varies between sensors, such a multiple-sensor system can be viewed as samples of a scale-space signal representation. It is demonstrated that if the spatial transfer function of the sensors are Gaussian, then scale-space filtering can be used to recover small-scale (fine-resolution) information through extrapolation in scale. As an example of multiple-sensor processing, multispectral processing of remote sensing, in which images of surface scenes are simultaneously generated at different (center) frequencies, is considered. The fingerprints of extrapolated signals approximate the actual multispectral fingerprints at small scales and can be used when the multispectral fingerprints are not available Thesaurus computerised pattern recognition; computerised picture processing ; filtering and prediction theory; remote sensing Other Terms fine resolution information recovery; Gaussian spatial transfer function; surface scene images; scale-space filtering; multiple-sensor processing; multispectral processing; remote sensing; extrapolated signals; multispectral fingerprints ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3875726 AbstractNos. B91033569; C91026462 References 11 U.S. Copyright Clearance Center Code CH2847-2/90/0000-2033$01.00 Country Pub. USA date 1174 ------------------------------------------------------------ Author Cung, H.M.; Cohen, P.; Boulanger, P.; Dept. of Electr. Eng., Ecole Polytech. de Montreal, Que., Canada Title Multiscale edge detection and classification in range images Source Proceedings 1990 IEEE International Conference on Robotics and Automation (Cat. No.90CH2876-1); Part: Cincinnati, OH, USA; Part: 13-18 May 1990; Sponsored by: IEEE; Los Alamitos, CA, USA; IEEE Comput. Soc. Press; 3 vol. xxxii+2184; 1990; pp. 2038-44 vol.3 Abstract An edge detection and classification scheme for range images which produces a multiscale representation in terms of well- localized depth and orientation edges is presented. The extraction is accomplished by detecting the presence of significant edges at a coarse scale and then determining their precise location by tracking them over decreasing scale. An adaptive multiscale thresholding is applied during this focusing process ro inhibit the attraction of insignificant details. Once focused, the edges are classified into the categories of true edge and diffuse edge by invoking classification rules derived from a mathematical analysis of edge displacement and branching over scale-space. Experimental results illustrate the robustness of the approach in the presence of noise and its performance with synthetic and real images of varying complexity. Comparisons with recently published techniques point out the improved performance of the approach, especially when the images contain substantially overlapping objects Thesaurus pattern recognition; picture processing Other Terms multiscale edge detection; edge classification; localized depth edges; range images; multiscale representation; orientation edges; tracking; adaptive multiscale thresholding; edge displacement; branching; scale-space; noise ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3875316 AbstractNos. B91033543; C91026438 ISBN or SBN 0 8186 9061 5 References 16 U.S. Copyright Clearance Center Code CH2876-1/90/0000-2038$1.00 Country Pub. USA date 1175 ------------------------------------------------------------ Author de Ridder, H.; Majoor, G.M.M.; Inst. for Perception Res., Eindhoven, Netherlands Title Numerical category scaling: an efficient method for assessing digital image coding impairments Source Human Vision and Electronic Imaging: Models, Methods and Applications; Part: Santa Clara, CA, USA; Part: 12-14 Feb. 1990; Sponsored by: SPIE; SPSE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1249; 1990; pp. 65-77 Abstract A problem in perceptual image quality assessment is the evaluation of the visible effects of digital image coding on the perceptual quality of images displayed on video screens. These effects are anticipated to be too small to be assessed by the widely employed method of rating on a category scale consisting of adjectives. A possible solution to this problem is to enhance the flexibility of category scaling by using numbers instead of adjectives. Experiments are described in which numerical category scaling has been used to assess impairment of perceptual image quality due to quantization errors in scale-space coding. The results show that direct numerical category scaling is an efficient method for assessing slight effects like the ones usually encountered in digitally coded images; direct category scaling and a scaling procedure in accordance with functional measurement theory end in the same functional relationship between impairment and degree of quantization; and unrelated impairments add up to form the overall impression of impairment Thesaurus analogue-digital conversion; coding errors; encoding; picture processing; screens (display); visual perception Other Terms encoding; picture processing; visual perception; digital image coding; perceptual image quality assessment; video screens; quantization errors; scale-space coding; quantization; impairments ClassCodes A8732S; A4230V; A8732Q; B6140C; B6120B; B7260; B6430; B7500 Article Type Experimental Coden PSISDG Language English RecordType Conference ControlNo. 3866451 AbstractNos. A91056424; B91033467 ISSN 0277786X References 23 Country Pub. USA date 1172 ------------------------------------------------------------ Author Brunnstrom, K.; Eklundh, J.-O.; Lindeberg, T.; Dept. of Numerical Anal. & Comput. Sci., R. Inst. of Technol., Stockholm, Sweden Title On scale and resolution in active analysis of local image structure Source First European Conference on Computer Vision; Part: Antibes, France; Part: 23-27 April 1990; Sponsored by: INRIA; Image and Vision Computing; vol.8, no.4; Nov. 1990; pp. 289-96 Abstract Focus-of-attention is extremely important in human visual perception. If computer vision systems are to perform tasks in a complex, dynamic world they will have to be able to control processing in a way that is analogous to visual attention in humans. Problems connected to foveation (examination of selected regions of the world at high resolution) are examined. In particular, the problem of finding and classifying junctions from this aspect is considered. It is shown that foveation as simulated by controlled, active zooming in conjunction with scale- space techniques allows for robust detection and classification of junctions Thesaurus computer vision Other Terms scale; resolution; active analysis; local image structure; human visual perception; computer vision systems; foveation; active zooming; scale-space techniques; classification ClassCodes C1250; C5260B Article Type Practical; Theoretical / Mathematical Coden IVCODK Language English RecordType Conference ControlNo. 3856950 AbstractNos. C91026335 ISSN 02628856 References 16 U.S. Copyright Clearance Center Code 0262-8856/90/040289-08$3.00 Country Pub. UK date 1174 ------------------------------------------------------------ Author Martens, J.-B.; Inst. for Perception Res., Eindhoven, Netherlands Title Application of scale space to image coding Source IEEE Transactions on Communications; IEEE Trans. Commun. (USA); vol.38, no.9; Sept. 1990; pp. 1585-91 Abstract The continuous formulation of scale space is briefly reviewed. It is shown that deriving a discrete formulation of scale space requires the solution to a more general problem; the optimum approximation of a signal by local patterns. The consequences of the theory for the Laplacian image pyramid are discussed. A pyramid coding scheme based on the discrete scale-space formulation is derived. Preliminary coding results on real images are presented. Down to 1 b/pixel, the quality of the coded images is usually very close to that of the originals. Bit rates below 0. 5 b/pixel imply a too coarse quantization, or even deletion, of the prediction error image at the smallest scale and, consequently, always result in images that are noticeably unsharp. In the intermediate region, different degrees of quantization noise and unsharpness are present. At comparable data rates, the linear variation coder generates less quantization noise in uniform regions, while the scale-space coder gives a slightly better edge reproduction Thesaurus encoding; filtering and prediction theory; picture processing Other Terms scale space filtering; image quality; image coding; Laplacian image pyramid; pyramid coding scheme; discrete scale-space formulation ClassCodes B6140C; B6120B; C1250 Article Type Theoretical / Mathematical Coden IECMBT Language English RecordType Journal ControlNo. 3848128 AbstractNos. B91026543; C91020007 ISSN 00906778 References 18 U.S. Copyright Clearance Center Code 0090-6778/90/0900-1585$01.00 Country Pub. USA date 1179 ------------------------------------------------------------ Author Etoh, M.; Tomono, A.; Kishino, F.; ATR Commun. Syst. Res. Lab., Kyoto, Japan Title Stereo-based description by generalized cylinder complexes from occluding contours Source Transactions of the Institute of Electronics, Information and Communication Engineers D-II; Trans. Inst. Electron. Inf. Commun. Eng. D-II (Japan); vol.J73D-II, no.9; Sept. 1990; pp. 1402-12 Abstract Three-dimensional description of a straight homogeneous generalized cylinder (SHGC) using axis-based stereo, and a contour line segmentation method are described. In this approach, stereo contour images are used. A pair of line segments are extracted assuming that they are the extremal contour of an SHGC. The axis of the SHGC's contour image is determined as smoothed local symmetry (SLS) of the line segments. The SHGC's axis is recovered by the stereo match of axes of SLS. For the line segmentation, an interval tree structure is built taking feature points as the curvature extrema using scale-space analysis. As a result, actual cylindrical objects are described as three- dimensional axes Thesaurus computer vision; computerised pattern recognition Other Terms generalized cylinder complexes; occluding contours; straight homogeneous generalized cylinder; axis-based stereo; contour line segmentation method; stereo contour images; extremal contour; smoothed local symmetry; interval tree structure; feature points; curvature extrema; scale-space analysis; cylindrical objects; three-dimensional axes ClassCodes C5260B; C1250 Article Type Theoretical / Mathematical; Experimental Coden DTGDE7 Language Japanese RecordType Journal ControlNo. 3816935 AbstractNos. C91016062 References 19 Country Pub. Japan date 1179 ------------------------------------------------------------ Author Schenk, T.; Jin-Cheng Li; Toth, C.K.; Dept. of Geodetic Sci. & Surveying, Ohio State Univ., Columbus, OH, USA Title Hierarchical approach to reconstruct surfaces by using iteratively rectified imagery Source Close-Range Photogrammetry Meets Machine Vision; Part: Zurich, Switzerland; Part: 3-7 Sept. 1990; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1395, pt.1; 1990; pp. 464-70 Abstract A new approach to reconstruct the three-dimensional surface of the object space from digital image is described. All the object points obtained by an automatic orientation procedure lead to a first approximation of the surface. Edges are computed for one image and matched to the other image by grey level correlation or least-squares matching through the scale space. To every discrete step in the scale space there exists the digital stereopair (image pyramid), the corresponding surface (digital elevation model DEM) and the warped images. The warped images in this discrete scale space representation correspond to digital orthophotos obtained from the DEMs that result from matching the image pyramid. The authors propose to use the warped images on every successive level in the image pyramid in order to reduce the foreshortening problems associated with any area-based matching method. The method and some experimental results are reported Thesaurus computerised picture processing; iterative methods; least squares approximations Other Terms hierarchical system; iteratively rectified imagery; three- dimensional surface; object space; digital image; automatic orientation procedure; grey level correlation; least-squares matching; digital stereopair; image pyramid; digital elevation model; warped images; digital orthophotos; area-based matching ClassCodes A4230V; A0260; B6140C; B0290F; C5260B; C7410H; C4130 Article Type Practical; Experimental Coden PSISDG Language English RecordType Conference ControlNo. 3813144 AbstractNos. A91022158; B91011222; C91016025 ISSN 0277786X References 9 Country Pub. USA date 1179 ------------------------------------------------------------ Author Lindeberg, T.; Eklundh, J.-O.; Comput. Vision & Active Perception Lab., Inst. of Technol., Stockholm, Sweden Title Construction of a scale-space primal sketch Source BMVC90 Proceedings of the British Machine Vision Conference; Part: Oxford, UK; Part: 24-27 Sept. 1990; Oxford, UK; BMVC 90; xxiii+426; 1990; pp. 97-102 Abstract The authors present a multi-scale representation of grey-level shape, called scale-space primal sketch that makes explicit features in scale-space as well as the relations between features at different levels of scale. The representation gives a qualitative description of the image structure that allows for extraction of significant image structure-stable scales and regions of interest-in a solely bottom-up data-driven manner. Hence, it can be seen as preceding further processing, which can then be properly tuned. Experiments on real imagery demonstrate that the proposed theory gives perceptually intuitive results Thesaurus computer vision; computerised pattern recognition; computerised picture processing Other Terms scale space representation; multi-scale representation; grey- level shape; scale-space primal sketch; image structure; real imagery ClassCodes B6140C; C5260B; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3811153 AbstractNos. B91011311; C91009872 References 11 Country Pub. UK date 1179 ------------------------------------------------------------ Author Wilson, R.; spann, M.; Warwick Univ., Coventry, UK Title A new approach to clustering Source Pattern Recognition; Pattern Recognit. (UK); vol.23, no.12; 1990; pp. 1413-25 Abstract Estimation theory is used to derive a new approach to the clustering problem. The new method is a unification of centroid and mode estimation, achieved by considering the effect of spatial scale on the estimator. The result is a multiresolution method which spans a range of spatial scales, giving enhanced robustness both to noise in the data and to changes of scale in the data, by using comparison between scales as a test of cluster validity. Iterative and non-iterative algorithms based on the new estimator are presented and are shown to be more accurate than simple scale-space filtering in identifying and locating the cluster centres from noisy test data. Results from a wide range of applications are used to illustrate the power and versatility of the new method Thesaurus estimation theory; iterative methods; pattern recognition Other Terms pattern recognition; iterative algorithm; noniterative algorithm ; clustering; spatial scale; multiresolution method ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 3807536 AbstractNos. B91011194; C91007695 ISSN 00313203 References 21 U.S. Copyright Clearance Center Code 0031-3203/90/$3.00+.00 Country Pub. UK date 1171 ------------------------------------------------------------ Author Ueda, N.; Suzuki, S.; NTT Human Interface Labs., Yokosuka, Japan Title A matching algorithm of deformed planar curves using multiscale convex/concave structures Source Transactions of the Institute of Electronics, Information and Communication Engineers D-II; Trans. Inst. Electron. Inf. Commun. Eng. D-II (Japan); vol.J73D-II, no.7; July 1990; pp. 992-1000 Abstract This paper proposes a multiscale segment matching method based on hierarchical convex/concave structures. First, multiscale representations are obtained using curvature scale space filtering and then inflection point correspondences are computed between consecutive smoothed shapes. This correspondence is useful for decreasing data space and calculation time of the scale space descriptions. Then the authors define the dissimilarity between two multiscale segments and optimize total dissimilarities between two multiscale shapes using a dynamic programming technique. Therefore, this method, unlike conventional methods, is applicable to the matching of highly deformed shapes. Experimental results show the usefulness of the proposed method Thesaurus computerised pattern recognition; computerised picture processing Other Terms deformed planar curves; multiscale segment matching; hierarchical convex/concave structures; multiscale representations; curvature scale space filtering; inflection point correspondences; consecutive smoothed shapes; scale space descriptions; dissimilarity; multiscale segments; dynamic programming; deformed shapes ClassCodes C5260B; C1250 Article Type Theoretical / Mathematical; Experimental Coden DTGDE7 Language Japanese RecordType Journal ControlNo. 3795991 AbstractNos. C91009812 References 4 Country Pub. Japan date 1177 ------------------------------------------------------------ Author Topkar, V.; Kjell, B.; Sood, A.; George Mason Univ., Fairfax, VA, USA Title Object detection using scale-space Source Applications of Artificial Intelligence VIII; Part: Orlando, FL, USA; Part: 17-19 April 1990; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1293, pt.1; 1990; pp. 2-13 Abstract Scale-space representation is a topic of active research in computer vision. Most of the work so far has concentrated on image reconstruction from the scale-space representation. In this paper the authors discuss the use of scale-space representation for object detection. They have proposed a model-based approach and have developed an algorithm to implement it. Channel integration is the heart of the algorithm and there are a number of unresolved issues in it. Object detection is possible only if the objects of interest are different from the noise and clutter in certain features. The authors have used two different images, one with good signal to noise ratio and the other with poor signal to noise ratio. In the first image the distinguishing feature of the object is its signal strength and in the second image it is its size. Accordingly, the authors have studied two approaches to the channel integration: (i) based on the contrast value and (ii) based on edge focusing and splitting. The results of both approaches are presented and discussed Thesaurus computer vision; spatial reasoning Other Terms object size; image reconstruction; scale-space representation; object detection; model-based approach; clutter; signal to noise ratio; signal strength; channel integration; contrast; edge focusing; splitting ClassCodes C1250 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3791472 AbstractNos. C91007706 ISSN 0277786X References 13 Country Pub. USA date 1174 ------------------------------------------------------------ Author Young Won Lim; Sang Uk Lee; Dept. of Control & Instrumentation Eng., Seoul Nat. Univ., South Korea Title On the color image segmentation algorithm based on the thresholding and the fuzzy c-means techniques Source Pattern Recognition; Pattern Recognit. (UK); vol.23, no.9; 1990; pp. 935-52 Abstract A segmentation algorithm for color images based on the thresholding and the fuzzy c-means (FCM) techniques is presented. The scale-space filter is used as a tool for analyzing the histograms of three color components. The methodology uses a coarse-fine concept to reduce the computational burden required for the FCM. The coarse segmentation attempts to segment coarsely using the thresholding technique, while the fine segmentation assigns the pixels, which remain unclassified after the coarse segmentation, to the closest class using the FCM. Attempts also have been made to compare the performance of the proposed algorithm with other existing algorithms-Ohlander's, Rosenfeld's, and Bezdek's. Intensive computer simulation has been performed and the results are discussed in this paper Thesaurus computerised picture processing; filtering and prediction theory; fuzzy set theory Other Terms computerised picture processing; color image segmentation algorithm; thresholding; fuzzy c-means; scale-space filter; coarse-fine concept ClassCodes B6140C; B0250; C1250; C5260B; C1160 Article Type Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 3785592 AbstractNos. B91004036; C91000884 ISSN 00313203 References 23 U.S. Copyright Clearance Center Code 0031-3203/90/$3.00+.00 Country Pub. UK date 1171 ------------------------------------------------------------ Author Vaezi, M.; Bavarian, B.; Dept. of Electr. & Comput. Eng., California Univ., Irvine, CA, USA Title Composite-scale Gaussian filtering Source Conference Record. Twenty-Third Asilomar Conference on Signals, Systems ands Computers (IEEE Cat. No.89-CH2836-5); Part: Pacific Grove, CA, USA; Part: 30 Oct.-1 Nov. 1989; Sponsored by: IEEE; Naval Postgraduate Sch.; San Jose State Univ; San Jose, CA, USA; Maple Press; 2 vol. xix+1064; 1989; pp. 741-3 vol.2 Editor Chen, R.R. Abstract An approach to the scale-space Gaussian filtering is described. This technique involves the use of a nonlinear function of input intensity and its first (and second) derivative(s) for the scale of the Gaussian filter where image smoothing and the correct space location of the zero crossings in an image are detected in one pass of convolution. The possible mathematical structure describing the optimality of this method is discussed. Illustrative examples are presented Thesaurus computerised picture processing; digital filters; filtering and prediction theory Other Terms input intensity nonlinear function; digital image filters; scale-space Gaussian filtering; image smoothing; correct space location; zero crossings; convolution ClassCodes B6140; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3776256 AbstractNos. B90078005; C91001081 References 6 U.S. Copyright Clearance Center Code 23ACSSC-12/89/0741$1.00 Country Pub. USA date 1167 ------------------------------------------------------------ Author Acharya, R.; Electr. & Comput. Eng., State Univ. of New York, Buffalo, NY, USA Title Multidimensional image analysis and mathematical morphology Source Sixth Multidimensional Signal Processing Workshop (Cat. No. 89TH0290-7); Part: Pacific Grove, CA, USA; Part: 6-8 Sept. 1989; Sponsored by: IEEE; New York, NY, USA; IEEE; 242; 1989; pp. 203 Abstract Summary form only given. Multidimensional operators based on mathematical morphology have been proposed for image segmentation. Mathematical morphology is basically a set theory. It provides the concept of a structuring element to probe the image with arbitrary geometric patterns, in order to capture the topological properties of the image. The classical operators have been extended to multidimensions. A morphological approach to scale- space filtering has been developed. Multiscale morphological openings that nonlinearly smooth the image without blurring the features (edges) have been used. The approach has been formulated within the framework of alternating sequential filters (ASF) Thesaurus filtering and prediction theory; picture processing; set theory Other Terms multiscale morphological openings; multidimensional operators; mathematical morphology; image segmentation; set theory; arbitrary geometric patterns; topological properties; scale- space filtering; alternating sequential filters ClassCodes B6140C; B0250; C1250; C1160 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3763813 AbstractNos. B90078185; C90067795 References 0 Country Pub. USA date 1166 ------------------------------------------------------------ Author Perona, P.; Dept. of Electr. Eng. & Comput. Sci., California Univ., Berkeley, CA, USA Title Anisotropic diffusion processes in early vision Source Sixth Multidimensional Signal Processing Workshop (Cat. No. 89TH0290-7); Part: Pacific Grove, CA, USA; Part: 6-8 Sept. 1989; Sponsored by: IEEE; New York, NY, USA; IEEE; 242; 1989; pp. 68 Abstract Summary form only given. Images often contain information at a number of different scales of resolution, so that the definition and generation of a good scale space is a key step in early vision. A scale space in which object boundaries are respected and smoothing only takes place within these boundaries has been defined that avoids the inaccuracies introduced by the usual method of low-pass-filtering the image with Gaussian kernels. The new scale space is generated by solving a nonlinear diffusion differential equation forward in time (the scale parameter). The original image is used as the initial condition, and the conduction coefficient c(x, y, t) varies in space and scale as a function of the gradient of the variable of interest (e.g. the image brightness). The algorithms are based on comparing the local values of different diffusion processes running in parallel on the same image Thesaurus picture processing Other Terms analog networks; digital architectures; edge detection; image compression; parallel computation structure; anisotropic diffusion process; early vision; scale space; object boundaries ; nonlinear diffusion differential equation; initial condition; conduction coefficient; gradient; image brightness; algorithms ClassCodes B6140C Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3763726 AbstractNos. B90078151 References 0 Country Pub. USA date 1166 ------------------------------------------------------------ Author Besslich, P.W.; Forgber, E.; Dept. of Electr. Eng., Bremen Univ., West Germany Title Model-based 2D edge detection using bottom-up strategy Source Sixth Multidimensional Signal Processing Workshop (Cat. No. 89TH0290-7); Part: Pacific Grove, CA, USA; Part: 6-8 Sept. 1989; Sponsored by: IEEE; New York, NY, USA; IEEE; 242; 1989; pp. 30-1 Abstract Summary form only given. It has been shown that a bottom-up strategy using an improved version of the optimized 2D edge filter fits the requirements of edge detection in real-world 2D images better than the top-down approach. The results are obtained by first operating on the full resolution, gradually restricting it to improve detection capabilities. The 2D filter has been applied to a bottom-up multiresolution edge detection scheme. All contour segments are registered. The behavior is observed while a fine-to-coarse tracing is performed, and the segments are classified as to whether they carry relevant information or not. The optimal edge detector has been tested using two strategies: a top-down procedure and the bottom-up scale-space scheme. While the computational burden is the same in both cases, the bottom-up approach provides considerable improvement in noise suppression and always detects the full length of a contour as obtained at the finest scale Thesaurus computerised pattern recognition; computerised picture processing ; filtering and prediction theory; interference suppression Other Terms optimized 2D edge filter; bottom-up multiresolution edge detection scheme; contour segments; fine-to-coarse tracing; top-down procedure; bottom-up scale-space scheme; noise suppression ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3763708 AbstractNos. B90078144; C90067768 References 3 Country Pub. USA date 1166 ------------------------------------------------------------ Author Zuerndorfer, B.; Wakefield, G.H.; Dept. of Electr. Eng. & Comput. Sci., Michigan Univ., Ann Arbor, MI, USA Title Extensions of scale-space filtering to machine-sensing systems Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.12, no.9; Sept. 1990; pp. 868-82 Abstract Major components of scale-space theory are Gaussian filtering, and the use of zero-crossing thresholders and Laplacian operators. Properties of scale-space filtering may be useful for data analysis in multiresolution machine-sensing systems. However, these systems typically violate the Gaussian filter assumption, and often the data analyses used (e.g. trend analysis and classification) are not consistent with zero-crossing thresholders and Laplacian operators. The authors extend the results of scale-space theory to include these more general conditions. In particular, it is shown that relaxing the requirement of linear scaling allows filters to have non-Gaussian spatial characteristics, and that relaxing of the scale requirements (s to 0) of the impulse response allows the use of scale-space filters with limited frequency support (i.e. bandlimited filters). Bandlimited scale-space filters represent an important extension of scale-space analysis for machine sensing Thesaurus filtering and prediction theory; pattern recognition; picture processing Other Terms image sensing; machine vision; scale-space filtering; data analysis; multiresolution machine-sensing systems; zero- crossing thresholders; Laplacian operators ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3758654 AbstractNos. B90078048; C90067694 ISSN 01628828 References 18 U.S. Copyright Clearance Center Code 0162-8828/90/0900-0868$01.00 Country Pub. USA date 1179 ------------------------------------------------------------ Author Morita, S.; Kawashima, T.; Aoki, Y.; Fac. of Eng., Hokkaido Univ., Sapporo, Japan Title Pattern matching of 2-D shape using hierarchical description Source Transactions of the Institute of Electronics, Information and Communication Engineers D-II; Trans. Inst. Electron. Inf. Commun. Eng. D-II (Japan); vol.J73D-II, no.5; May 1990; pp. 717-27 Abstract Introduces a hierarchical description method of two-dimensional shapes suited for pattern matching. In the method, a shape is described as a string of scale-filtered line segments called the basic codon combination and is structured as a tree of segments along coarse-to-fine scale-space analysis. With the description, a compact database is created for similar shapes, because trees overlap at the coarse level. Using the database, analysis and matching of the tree structure is efficiently performed in a top- down form. Experimental results show that the system recognizes the target shape efficiently even if a shape is skewed or occluded Thesaurus computerised pattern recognition; database management systems; trees (mathematics) Other Terms top-down analysis; skewed shapes; occluded shapes; hierarchical description; two-dimensional shapes; pattern matching; string; scale-filtered line segments; basic codon combination; tree; coarse-to-fine scale-space analysis; compact database ClassCodes C1250; C5530 Article Type Practical; Experimental Coden DTGDE7 Language Japanese RecordType Journal ControlNo. 3749821 AbstractNos. C90067737 References 15 Country Pub. Japan date 1175 ------------------------------------------------------------ Author Saund, E.; Xerox Palo Alto Res. Center, CA, USA Title Symbolic construction of a 2-D scale-space image Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.12, no.8; Aug. 1990; pp. 817-30 Abstract A symbolic approach to constructing a multiscale primitive shape description to 2-D binary (silhouette) shape images is presented. In contrast to contour or region smoothing techniques, grouping operations are performed over collections of tokens residing on a scale-space blackboard. Two types of grouping operations are identified that, respectively, aggregate edge primitives at one scale into edge primitives at a coarser scale and group edge primitives into partial-region assertions, including curved contours, primitive corners, and bars. Procedures to perform these computations are presented Thesaurus pattern recognition; picture processing Other Terms pattern recognition; picture processing; 2-D scale-space image; multiscale primitive shape description; grouping operations; edge primitives; partial-region assertions; curved contours; primitive corners; bars ClassCodes B6140C; C1250 Article Type Practical Coden ITPIDJ Language English RecordType Journal ControlNo. 3746886 AbstractNos. B90071185; C90067691 ISSN 01628828 References 35 U.S. Copyright Clearance Center Code 0162-8828/90/0800-0817$01.00 Country Pub. USA date 1178 ------------------------------------------------------------ Author Jepson, A.D.; Fleet, D.J.; Dept. of Comput. Sci., Toronto Univ., Ont., Canada Title Scale-space singularities Source Computer Vision - ECCV 90. First European Conference on Computer Vision Proceedings; Part: Antibes, France; Part: 23-27 April 1990; Sponsored by: INRIA; Berlin, West Germany; Springer-Verlag; xii+618; 1990; pp. 50-5 Editor Faugeras, O. Abstract Phase-based techniques for the measurement of binocular disparity and image velocity are encouraging, especially because of the stability of band-pass phase information with respect to deviations from image translation that are typical in projections of 3-D scenes. Despite this stability, phase is unreliable in the neighbourhoods of phase singularities. This instability is described, and it is shown that singularity neighbourhoods may be detected using simple constraints on the local frequency and the amplitude of the filter output. Finally, these results are discussed briefly in the context of binocular disparity measurement Thesaurus computer vision; computerised picture processing Other Terms phase-based techniques; image velocity measurement; 3D scene projections; stability; band-pass phase information; deviations ; image translation; phase singularities; instability; singularity neighbourhoods; local frequency; amplitude; filter output; binocular disparity measurement ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 3742851 AbstractNos. B90071249; C90061261 ISBN or SBN 3 540 52522 X References 12 Country Pub. West Germany date 1174 ------------------------------------------------------------ Author Perona, P.; Malik, J.; Dept. of Electr. Eng. & Comput. Sci., California Univ., Berkeley, CA, USA Title Scale-space and edge detection using anisotropic diffusion Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.12, no.7; July 1990; pp. 629-39 Abstract A new definition of scale-space is suggested, and a class of algorithms used to realize a diffusion process is introduced. The diffusion coefficient is chosen to vary spatially in such a way as to encourage intraregion smoothing rather than interregion smoothing. It is shown that the 'no new maxima should be generated at coarse scales' property of conventional scale space is preserved. As the region boundaries in the approach remain sharp, a high-quality edge detector which successfully exploits global information is obtained. Experimental results are shown on a number of images. Parallel hardware implementations are made feasible because the algorithm involves elementary, local operations replicated over the image Thesaurus filtering and prediction theory; pattern recognition; picture processing Other Terms parallel processing; picture processing; edge detection; anisotropic diffusion; scale-space; intraregion smoothing ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical; Experimental Coden ITPIDJ Language English RecordType Journal ControlNo. 3735403 AbstractNos. B90071174; C90061167 ISSN 01628828 References 22 U.S. Copyright Clearance Center Code 0162-8828/90/0700-0629$01.00 Country Pub. USA date 1177 ------------------------------------------------------------ Author Brunnstrom, K.; Eklundh, J.-O.; Lindeberg, T.; Comput, Vision & Associative Pattern Process. Lab., R. Inst. of Technol., Stockholm, Sweden Title On scale and resolution in the analysis of local image structure Source Computer Vision - ECCV 90. First European Conference on Computer Vision Proceedings; Part: Antibes, France; Part: 23-27 April 1990; Sponsored by: INRIA; Berlin, West Germany; Springer-Verlag; xii+618; 1990; pp. 3-12 Editor Faugeras, O. Abstract Focus-of-attention is extremely important in human visual perception. If computer vision systems are to perform tasks in a complex, dynamic world they will have to be able to control processing in a way that is analogous to visual attention in humans. The paper investigates problems in connection with foveation, that is examining selected regions of the world at high resolution. It considers the problem of finding and classifying junctions from this aspect. It shows that foveation as simulated by controlled, active zooming in conjunction with scale-space techniques allows robust detection and classification of junctions Thesaurus computer vision; computerised picture processing Other Terms focus of attention; robust classification; scale; resolution; local image structure; human visual perception; computer vision systems; visual attention; foveation; selected regions; world; high resolution; junctions; active zooming; scale-space; robust detection ClassCodes C5260B Article Type Practical Language English RecordType Conference ControlNo. 3731228 AbstractNos. C90063685 ISBN or SBN 3 540 52522 X References 15 Country Pub. West Germany date 1174 ------------------------------------------------------------ Author Wada, T.; Yi He Gu; Sato, M.; Res. Lab. of Precision Machinery & Electron., Tokyo Inst. of Technol., Yokohama, Japan Title Scale-space filtering for periodic waveforms Source Transactions of the Institute of Electronics, Information and Communication Engineers D-II; Trans. Inst. Electron. Inf. Commun. Eng. D-II (Japan); vol.J73D-II, no.4; April 1990; pp. 544-52 Abstract Scale-space filtering is a multi-resolution filtering technique, which has suitable properties for a hierarchical description of waveform. The most important property is that the number of zero crossings will never increase, while increasing the scale- parameter. It is called the monotonicity of zero crossings. The authors construct the scale-space filtering for periodic waveforms to hold the monotonicity. The derived filtering kernel is the elliptic theta function of the third kind I/sub 3/. The basic properties of this filtering is cleared. Some examples of structural analysis are shown, and the basic differences between periodic scale-space filtering and non-periodic are discussed Thesaurus filtering and prediction theory; waveform analysis Other Terms hierarchical waveform description; multi-resolution filtering; zero crossings; periodic waveforms; elliptic theta function ClassCodes B6140; C1260 Article Type Theoretical / Mathematical Coden DTGDE7 Language Japanese RecordType Journal ControlNo. 3727560 AbstractNos. B90063596; C90061446 References 8 Country Pub. Japan date 1174 ------------------------------------------------------------ Author Etoh, M.; Tomono, A.; Kobayashi, Y.; ATR Commun. Syst. Res. Lab., Kyoto, Japan Title Cylindrical part recognition in occluding contours Source Intelligent Robots and Computer Vision VIII: Algorithms and Techniques; Part: Philadelphia, PA, USA; Part: 6-10 Nov. 1989; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1192, pt.1; 1990; pp. 353-62 Abstract Three-dimensional reconstruction method of a straight homogeneous generalized cylinder model using 'axis-based stereo', and contour line segmentation method are described. To achieve the axis-based stereo matching, stereo contour images are used. In each contour image, a pair of contour line segments are assumed to be the extremal contour of a cylinder and are interpreted as a 'ribbon'. A pair of ribbons over the stereo contour images are interpreted as a 'cylinder'. The cylinder's axis is determined by the stereo match of two ribbon's axes in space. For the line segmentation, an interval tree structure is built taking feature points as the curvature extrema using scale-space analysis, and splitting the feature points' interval recursively to satisfy a line regularity Thesaurus computer vision; computerised pattern recognition; computerised picture processing Other Terms 3D image reconstruction; occluding contour recognition; computerised picture processing; pattern recognition; cylinder model; contour line segmentation; stereo matching; tree structure; feature points; curvature; scale-space analysis ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3698298 AbstractNos. B90056884; C90055154 ISSN 0277786X References 18 Country Pub. USA date 1168 ------------------------------------------------------------ Author Meer, P.; Sher, C.A.; Rosenfeld, A.; Center for Autom. Res., Maryland Univ., College Park, MD, USA Title The chain pyramid: hierarchical contour processing Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.12, no.4; April 1990; pp. 363-76 Abstract A novel hierarchical approach toward fast parallel processing of chain-codable contours is presented. The environment, called the chain pyramid, is similar to a regular nonoverlapping image pyramid structure. The artifacts of contour processing on pyramids are eliminated by a probabilistic allocation algorithm. Building of the chain pyramid is modular, and for different applications new algorithms can be incorporated. Two applications are described: smoothing of multiscale curves and gap bridging in fragmented data. The latter is also employed for the treatment of branch points in the input contours. A preprocessing module allowing the application of the chain pyramid to raw edge data is also described. The chain pyramid makes possible fast, O(log(image/sub -/size)), computation of contour representation in discrete scale-space Thesaurus computerised picture processing; parallel processing Other Terms computerised picture processing; chain pyramid; hierarchical contour processing; parallel processing; chain-codable contours; probabilistic allocation algorithm; smoothing; multiscale curves; gap bridging; fragmented data; raw edge data ClassCodes C5260B Article Type Practical; Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3663309 AbstractNos. C90046136 ISSN 01628828 References 33 U.S. Copyright Clearance Center Code 0162-8828/99/0400-0363$01.00 Country Pub. USA date 1174 ------------------------------------------------------------ Author Siebert, J.P.; Urquhart, C.W.; BBN Syst. & Technol., Heriot-Watt Res. Park, Edinburgh, UK Title Active stereo: texture enhanced reconstruction Source Electronics Letters; Electron. Lett. (UK); vol.26, no.7; 29 March 1990; pp. 427-30 Abstract An experiment is presented that combines active and passive vision techniques. The authors have demonstrated that it is possible to recover dense range map information from stereograms of scenes bathed in random noise 'textured light' by employing scale-space signal matching techniques Thesaurus computer vision; computerised picture processing; texture Other Terms active stereo; CCD camera; random noise textured light; multiscale signed matches; texture enhanced reconstruction; passive vision; dense range map information; stereograms; scale-space signal matching techniques ClassCodes A4230V; B6140C; C1250 Article Type Experimental Coden ELLEAK Language English RecordType Journal ControlNo. 3652112 AbstractNos. A90080563; B90045575; C90038342 ISSN 00135194 References 3 U.S. Copyright Clearance Center Code 0013-5194/90/$3.00+0.00 Country Pub. UK date 1173 ------------------------------------------------------------ Author Lindeberg, T.; Comput. Vision & Associative Pattern Process. Lab., R. Inst. of Technol., Stockholm, Sweden Title Scale-space for discrete signals Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.12, no.3; March 1990; pp. 234-54 Abstract A basic and extensive treatment of discrete aspects of the scale- space theory is presented. A genuinely discrete scale-space theory is developed and its connection to the continuous scale- space theory is explained. Special attention is given to discretization effects, which occur when results from the continuous scale-space theory are to be implemented computationally. The 1D problem is solved completely in an axiomatic manner. For the 2D problem, the author discusses how the 2D discrete scale space should be constructed. The main results are as follows: the proper way to apply the scale-space theory to discrete signals and discrete images is by discretization of the diffusion equation, not the convolution integral; the discrete scale space obtained in this way can be described by convolution with the kernel, which is the discrete analog of the Gaussian kernel, a scale-space implementation based on the sampled Gaussian kernel might lead to undesirable effects and computational problems, especially at fine levels of scale; the 1D discrete smoothing transformations can be characterized exactly and a complete catalogue is given; all finite support 1D discrete smoothing transformations arise from repeated averaging over two adjacent elements (the limit case of such an averaging process is described); and the symmetric 1D discrete smoothing kernels are nonnegative and unimodal, in both the spatial and the frequency domain Thesaurus discrete systems; signal processing Other Terms signal processing; nonnegative kernels; unimodal kernels; spatial domain; discrete signals; discrete scale-space theory; diffusion equation; Gaussian kernel; discrete smoothing transformations; frequency domain ClassCodes C1260 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3641638 AbstractNos. C90038539 ISSN 01628828 References 23 U.S. Copyright Clearance Center Code 0162-8828/90/0300-0234$01.00 Country Pub. USA date 1173 ------------------------------------------------------------ Author Leymarie, F.; Levine, M.D.; McGill Res. Center for Intelligent Mach., McGill Univ., Montreal, Que., Canada Title Shape features using curvature morphology Source Visual Communications and Image Processing IV; Part: Philadelphia, PA, USA; Part: 8-10 Nov. 1989; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1199, pt.1; 1989; pp. 390-401 pt.1 Abstract The authors briefly present a scheme for obtaining the discrete curvature function of planar contours based on the chain-code representation of a boundary. They propose a method for extracting important features from the curvature function such as extrema or peaks, and segments of constant curvature, using mathematical morphological operations on functions. On the basis of these morphological operations, they suggest a new scale-space representation for curvature named the Morphological Curvature Scale-Space. Advantages over the usual scale-space approaches are shown Thesaurus computer vision; curvature measurement Other Terms shape feature extraction; computer vision; curvature morphology; discrete curvature function; chain-code representation; mathematical morphological operations; scale-space representation ClassCodes C1250; C1160; C5260B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3617146 AbstractNos. C90031701 ISSN 0277786X References 34 Country Pub. USA date 1168 ------------------------------------------------------------ Author Hummel, R.; Moniot, R.; Courant Inst. of Math. Sci., New York Univ., NY, USA Title Reconstructions from zero crossings in scale space Source IEEE Transactions on Acoustics, Speech and Signal Processing; IEEE Trans. Acoust. Speech Signal Process. (USA); vol.37, no.12; Dec. 1989; pp. 2111-30 Abstract In computer vision, the one-parameter family of images obtained from the Laplacian-of-a-Gaussian-filtered version of the image, parameterized by the width of the Gaussian, has proved to be a useful data structure for the extraction of feature data. In particular, the zero crossings of this so-called scale-space data are associated with edges and have been proposed by D. Marr (1982) and others as the basis of a representation of the image data. The question arises as to whether the representation is complete and stable. The authors survey some of the studies and results related to these questions as well as several studies that attempt reconstructions based on this or related representations. They formulate a novel method for reconstruction from zero crossings in scale space that is based on minimizing equation error, and they present results showing that the reconstruction is possible but can be unstable. They further show that the method applies when gradient data along the zero crossings are included in the representation, and they demonstrate empirically that the reconstruction is then stable Thesaurus computer vision; computerised picture processing; filtering and prediction theory Other Terms image reconstruction; picture processing; zero crossings; scale space; computer vision; one-parameter family; images; Laplacian-of-a-Gaussian-filtered; data structure; feature data; equation error ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Coden IETABA Language English RecordType Journal ControlNo. 3606735 AbstractNos. B90030589; C90024515 ISSN 00963518 References 41 U.S. Copyright Clearance Center Code 0096-3518/89/1200-2111$01.00 Country Pub. USA date 1169 ------------------------------------------------------------ Author Parvin, B.; Medioni, G.; Dept. of Electr. Eng., Univ. of Southern California, Los Angeles, CA, USA Title A constraint satisfaction network for matching 3D objects Source IJCNN: International Joint Conference on Neural Networks (Cat. No. 89CH2765-6); Part: Washington, DC, USA; Part: 18-22 June 1989; Sponsored by: IEEE; New York, NY, USA; IEEE TAB Neural Network Committee; 2 vol. (790+646); 1989; pp. 281-6 vol.2 Abstract A new approach is presented for matching visible surfaces of 3D objects using a constraint satisfaction network. This in turn provides the necessary basis for volumetric reconstruction from multiple views. By matching, the authors mean both to establish correspondence between individual faces and to compute 3D transform that would bring one in correspondence with the other. Toward this goal, constraints at three different levels of complexities are specified to produce stable and coherent assignments. The constraint satisfaction is implemented as a Hopfield network with an appropriate energy functional and minimized using simulated annealing. The system extracts objects faces by computing their bounding contours with adaptive scale space filtering. This process identifies important surface features such as jumps or occluding boundaries and creases. The pointwise feature descriptors are then linked, and an attributed graph is derived to represent the object. The nodes in the graph represent geometric surface features, and the links in the graph represent the relationship between adjacent surfaces. The authors present results on real images Thesaurus filtering and prediction theory; neural nets; pattern recognition; transforms Other Terms pattern recognition; neural nets; 3D objects matching; constraint satisfaction network; volumetric reconstruction; multiple views; 3D transform; Hopfield network; simulated annealing; adaptive scale space filtering; jumps; occluding boundaries; creases; pointwise feature descriptors; attributed graph ClassCodes B6140C; B0230; C1250; C1260; C1130; C1230 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3601690 AbstractNos. B90023846; C90024628 References 23 Country Pub. USA date 1163 ------------------------------------------------------------ Author Blostein, D.; Ahuja, N.; Coordinated Sci. Lab., Illinois Univ., Urbana, IL, USA Title Shape from texture: integrating texture-element extraction and surface estimation Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.11, no.12; Dec. 1989; pp. 1233-51 Abstract A method is presented for identifying texture elements while simultaneously recovering the orientation of textured surfaces. A multiscale region detector, based on measurements in a Del /sup 2/G (Laplacian-of-Gaussian) scale space, is used to construct a set of candidate texture elements. True elements are selected from the set of candidate elements by finding the planar surface that best predicts the observed areas of the latter. Results are shown for a variety of natural textures, including waves, flowers, rocks, clouds, and dirt clods Thesaurus pattern recognition; picture processing Other Terms surface orientation estimation; texture element identification; Laplacian-of-Gaussian scale space; Del /sup 2/G scale space; texture-element extraction; multiscale region detector; waves; flowers; rocks; clouds; dirt clods ClassCodes C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3572673 AbstractNos. C90018327 ISSN 01628828 References 27 U.S. Copyright Clearance Center Code 0162-8828/89/1200-1233$01.00 Country Pub. USA date 1169 ------------------------------------------------------------ Author Koenderink, J.J.; Dept. of Med. & Physiol. Phys., Utrecht Univ., Netherlands Title A hitherto unnoticed singularity of scale-space Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.11, no.11; Nov. 1989; pp. 1222-4 Abstract A hitherto unnoticed singularity of scale space occurs only at isolated points in scale space. Thus it does not generically occur for single images, but it can occur occasionally in members of time sequences (say). It occurs at those critical points of the image at which the Laplacean of the illuminance vanishes (a nongeneric condition). The structure of scale space in an infinitesimal neighborhood of such a singularity is explored. The effect of the singularity of mappings between copies of an image at different levels of resolution is evaluated and checked with a numerical calculation Thesaurus pattern recognition; picture processing Other Terms picture processing; pattern recognition; hitherto unnoticed singularity; scale-space; time sequences; mappings ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3572670 AbstractNos. B90017119; C90018326 ISSN 01628828 References 6 U.S. Copyright Clearance Center Code 0162-8828/89/1100-1222$01.00 Country Pub. USA date 1168 ------------------------------------------------------------ Author Estola, K.-P.; Machine Autom. Lab., Tech. Centre of Finland, Tampere, Finland Title Multirate Gaussian scale-space filtering Source Eurospeech 89. European Conference on Speech Communication and Technology; Part: Paris, France; Part: 26-28 Sept. 1989; Sponsored by: Assoc. Belge des Acousticiens; Assoc. Recherche Cognitive; Comm. Eur. Communities; et al; Edinburgh, UK; CEP Consultants; 2 vol. (xxiii+636+xxi+721); 1989; pp. 625-8 vol.1 Editor Tubach, J.P.; Mariani, J.J. Abstract This paper proposes multirate signal processing methods for realizing Gaussian scale-space filtering. The author introduces new computationally efficient interpolated Gaussian scale-space filters. Also, the use of decimators together with interpolated Gaussian filters is considered. The proposed filtering methods require dramatically less computation than conventional Gaussian scale-space filtering especially in cases where the scale changes with an integer factor. The new filters are extremely efficient when the change in scale is an integer power of two. However, also more complex scaling factors such as square root 2 can be efficiently realized Thesaurus filtering and prediction theory; signal processing Other Terms multirate Gaussian scale-space filtering; signal processing methods; decimators ClassCodes B6140; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3557786 AbstractNos. B90009917; C90011235 References 5 Country Pub. UK date 1166 ------------------------------------------------------------ Author Sato, M.; Wada, T.; Res. Lab. of Precision Machinery & Electron., Tokyo Inst. of Technol., Japan Title Structure line of waveforms and its application Source Proceedings of 6th Scandinavian Conference on Image Analysis; Part : Oulu, Finland; Part: 19-22 June 1989; Sponsored by: Int. Assoc. Pattern Recognition; Oulu, Finland; Pattern Recognition Soc. Finland; 2 vol.(xx+1253); 1989; pp. 868-73 vol.2 Editor Pietikainen, M.; Roning, J. Abstract A structure line is a hierarchical representation of waveforms based on scale space filtering. Structure line has the same topological property as a ternary tree, and represents the hierarchy of convex and concave regions of the waveform. The authors discuss an application of the structure line to computer vision. One of the most basic and difficult problems of computer vision is the reconstruction of the 3-D object from the multi- viewpoint images. There may be the occlusion of the characteristic points between the images, then one can't find the matching pair. So, one should know the occluded regions between the images before the correspondence process. They investigated the relation between the morphological transition of the structure line of the observed images and the transition of the scene. Then the relation between the transitions of the structure line and the scene transitions are cleared Thesaurus computer vision; waveform analysis Other Terms 3D object reconstruction; convex regions; waveforms; scale space filtering; ternary tree; concave regions; computer vision ; morphological transition; scene transitions ClassCodes C1250; C1120 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3557471 AbstractNos. C90011057 References 5 Country Pub. Finland date 1163 ------------------------------------------------------------ Author Krozel, J.; Artificial Intelligence Center, Hughes Res. Labs., Malibu, CA, USA Title Planning with abstraction: map data feature extraction in scale- space Source Applications of Artificial Intelligence VII; Part: Orlando, FL, USA; Part: 28-30 March 1989; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1095, pt.1; 1989; pp. 25-35 Abstract Autonomous vehicles often perform navigation and path planning using hierarchical control systems. These systems separate high and low level reasoning through an abstraction of the planning problem. For reasoning about terrain information, the author presents a method of abstraction that retains the finest level of resolution while progressing through greater levels of abstraction. Abstraction arises from a continuum of Gaussian smoothed terrain surfaces, each smoothed surface describes the terrain at a different scale of abstraction. He refers to this continuum as scale-space. For each level of abstraction, important features can be extracted from land elevation data for planning purposes. He presents this abstraction method, a graph representation for retaining scale-space information, and examples of how features from various levels of abstraction influence planning at different levels of a hierarchical control system Thesaurus automatic guided vehicles; computer vision; hierarchical systems ; knowledge based systems Other Terms autonomous vehicles; abstraction; map data feature extraction in scale-space; navigation; path planning; hierarchical control systems; planning problem; reasoning; terrain information; Gaussian smoothed terrain surfaces; land elevation data; graph representation ClassCodes C5260B; C3360; C6170 Article Type Practical Coden PSISDG Language English RecordType Conference ControlNo. 3548555 AbstractNos. C90013567 ISSN 0277786X References 5 Country Pub. USA date 1160 ------------------------------------------------------------ Author Rueff, M.; Fraunhofer Inst. for Manuf., Eng. & Autom., Stuttgart, West Germany Title Scale space filtering and the scaling regions of fractals Source From Pixels to Features. Proceedings of a Workshop; Part: Bonas, France; Part: 22-27 Aug. 1988; Amsterdam, Netherlands; North-Holland; xiii+416; 1989; pp. 49-60 Editor Simon, J.C. Abstract Fractal dimensions are quantities which have been shown to be useful in the classification and segmentation of textures with scaling behaviour. The application of the concept of fractal analysis to the study of irregular structures is demonstrated by optical roughness measurements. Problems arising in the numerical determination of fractal dimensions are briefly mentioned. Scale space filtering techniques are suggested to overcome some of these problems which in particular are given with the detection of the limited scaling regions of natural textures Thesaurus filtering and prediction theory; fractals; pattern recognition Other Terms scale space filtering; image analysis; lacunarity; fractals; classification; segmentation; scaling behaviour; fractal analysis; irregular structures; optical roughness measurements; fractal dimensions; natural textures ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3545746 AbstractNos. B90010063; C90005324 References 16 Country Pub. Netherlands date 1152 ------------------------------------------------------------ Author de Vel, O.Y.; Thomas, P.; Dept. of Comput. Sci., Waikato Univ., Hamilton, New Zealand Title Multi-transputer system for speech spectrogram processing Source Centre for Advanced Technology in Telecommunications. Australian Transputer and OCCAM User Group Conference Proceedings; Part: Melbourne, Vic., Australia; Part: 6-7 July 1989; Sponsored by: Centre Adv. Technol.Telecommun; Melbourne, Vic., Australia; Centre for Adv. Technol. Telecommun; ii+193; 1989; pp. 29-39 Abstract A parallel processing system used for the low-level, near- neighbourhood processing of images is presented. The system consists of a loosely-coupled two-dimensional quad-connected mesh network for transputers. It is capable of executing both linear and nonlinear image processing functions. Example functions include linear convolution, median filtering, edge detection using the scale-space technique, and simulated annealing. The system environment includes downloading/uploading the image and filter masks, execution timing information, and image display. Near-linear speed-up performances have been observed for all image processing functions, the interprocessor communication time being the dominating limiting factor. The parallel image processing system has been used for the enhancement and segmentation of speech spectrograms Thesaurus computerised picture processing; parallel processing; speech analysis and processing Other Terms multi-transputer system; image downloading; image uploading; image enhancement; image segmentation; speech spectrogram processing; parallel processing system; near-neighbourhood processing; loosely-coupled two-dimensional quad-connected mesh network; image processing functions; linear convolution; median filtering; edge detection; scale-space technique; simulated annealing; system environment; filter masks; execution timing information; image display; speed-up performances; interprocessor communication time ClassCodes B6140C; B6130; C5260B; C5585 Article Type Practical Language English RecordType Conference ControlNo. 3533692 AbstractNos. B90003028; C90007013 References 9 Country Pub. Australia date 1164 ------------------------------------------------------------ Author Granum, E.; Christensen, H.I.; Inst. of Electron. Syst., Aalborg Univ., Denmark Title On principles of motion analysis in real time Source Image Processing II; Part: Hamburg, West Germany; Part: 19-21 Sept. 1988; Sponsored by: Eur. Phys. Soc.; Eur. Federation for Appl. Opt.; SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1027; 1989; pp. 113-20 Abstract A number of motion analysis methods are reviewed and evaluated with regard to dependency on supplementary processing and with regard to current potential for real time application. Problems of ambiguity and noise are considered. Image differencing and the low and high level token matching approaches are only methods considered realistic for real time operation. Aspects of higher level motion analysis are discussed, and the essence put into a scheme for a model driven approach to real time function. A high level token matching version of the model driven scheme has been implemented, and its basic performance was demonstrated on data with occlusions. Although under conditions of a constraint scenario, real time motion analysis on computer is feasible without the need for very sophisticated hardware. A scale space extension to the implementation was demonstrated, to provide a potential approach for description and analysis of composite motion patterns Thesaurus computer vision; computerised picture processing Other Terms low level token matching; image differencing; computerised picture processing; motion analysis; real time; high level token matching; model driven scheme; scale space extension; composite motion patterns ClassCodes C5260B; C7410F Article Type Practical; Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3526180 AbstractNos. C90006901 ISSN 0277786X References 41 Country Pub. USA date 1153 ------------------------------------------------------------ Author Haglund, L.; Knutsson, H.; Granlund, G.H.; Comput. Vision Lab., Linkoping Univ., Sweden Title Scale analysis using phase representation Source Proceedings of 6th Scandinavian Conference on Image Analysis; Part : Oulu, Finland; Part: 19-22 June 1989; Sponsored by: Int. Assoc. Pattern Recognition; Oulu, Finland; Pattern Recognition Soc. Finland; 2 vol.(xx+1253); 1989; pp. 1118-25 vol.2 Editor Pietikainen, M.; Roning, J. Abstract Scale analysis and description has over the last years become one of the major research fields in image processing. There are two main reasons for this. A single filter has a particular limited pass band which may or may not be tuned to the different sized objects to be described. Secondly, size or scale is a descriptive feature in its own right. All of this requires the integration of measurements from different scales. The paper describes a new algorithm which detects the scale in which an event appears and disappears. In this way the scale space is subdivided into a number of intervals. Within each scale interval a consistency check is performed to get the certainty of the detection. The algorithms are shown to be simple operations if a continuous phase representation is used Thesaurus filtering and prediction theory; picture processing Other Terms scale analysis; scale description; phase representation; image processing; filter; measurements; algorithm; scale space; scale interval; consistency check; continuous phase representation ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3524361 AbstractNos. B90003018; C90000476 References 7 Country Pub. Finland date 1163 ------------------------------------------------------------ Author Lindeberg, T.; Comput. Vision & Associative pattern Processing Lab., R. Inst. of Technol., Stockholm, Sweden Title Scale-space for discrete images Source Proceedings of 6th Scandinavian Conference on Image Analysis; Part : Oulu, Finland; Part: 19-22 June 1989; Sponsored by: Int. Assoc. Pattern Recognition; Oulu, Finland; Pattern Recognition Soc. Finland; 2 vol.(xx+1253); 1989; pp. 1098-107 vol.2 Editor Pietikainen, M.; Roning, J. Abstract Addresses the formulation of a scale-space theory for one- dimensional discrete images. Two main subjects are treated. Which linear transformations remove structure in the sense that the number of local extrema (or zero-crossings) in the output image does not exceed the number of local extrema (or zero-crossings) in the original image? How should one create a multiresolution family of representations with the property that an image at a coarser level of scale never contains more structure than an image at a finer level of scale? The author proposes that there is only one reasonable way to define a scale-space for discrete images comprising a continuous scale parameter, namely by (discrete) convolution with the family of kernels T(n;t)=e/sup - t/I/sub n/(t), where I/sub n/ are the modified Bessel functions of integer order. Similar arguments applied in the continuous case uniquely lead to the Gaussian kernel Thesaurus picture processing Other Terms discrete images; scale-space theory; linear transformations; local extrema; zero-crossings; output image; continuous scale parameter; modified Bessel functions; Gaussian kernel ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3524359 AbstractNos. B90003016; C90000474 References 9 Country Pub. Finland date 1163 ------------------------------------------------------------ Author Mussigmann, U.; Fraunhofer Inst. for Manuf., Eng. & Autom., Stuttgart, West Germany Title Texture analysis, fractals and scale space filtering Source Proceedings of 6th Scandinavian Conference on Image Analysis; Part : Oulu, Finland; Part: 19-22 June 1989; Sponsored by: Int. Assoc. Pattern Recognition; Oulu, Finland; Pattern Recognition Soc. Finland; 2 vol.(xx+1253); 1989; pp. 987-94 vol.2 Editor Pietikainen, M.; Roning, J. Abstract The method of scale space filtering has been used till now in image analysis for the description and recognition of planar curves and two dimensional shapes. In this paper, the author presents a new method for the calculation of the fractal dimension of textures with the help of the scale space filtering. This fractal dimension is used as a quantitative measure for the classification and segmentation of textured images Thesaurus filtering and prediction theory; picture processing Other Terms picture processing; fractals; scale space filtering; image analysis; planar curves; textures; classification; segmentation ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3524344 AbstractNos. B90003001; C90000464 References 16 Country Pub. Finland date 1163 ------------------------------------------------------------ Author Kadirkamanathan, M.; Rayner, P.J.W.; Dept. of Eng., Cambridge Univ., UK Title A unified approach to on-line cursive script segmentation and feature extraction Source ICASSP-89: 1989 International Conference on Acoustics, Speech and Signal Processing (IEEE Cat. No.89CH2673-2); Part: Glasgow, UK; P art: 23-26 May 1989; Sponsored by: IEEE; New York, NY, USA; IEEE; 4 vol. 2833; 1989; pp. 1659-62 vol.3 Abstract A stroke segmentation algorithm based on curvature information processing and scale-space filtering is proposed. The strokes extracted are very close to strokes described by psychophysicists. A simple online author-dependent recognition system that uses the segmentation algorithm as a preprocessing stage is described and its performance evaluated for handwritten data obtained from five authors. The segmentation results indicate a performance much better than that of techniques based on direct estimation of penspeed or curvature. The scale-space plot of stroke boundaries also locates many descriptive features detectable in cursive script, allowing the programmer to choose any desired scale or scales. Once segmentation is performed, the problem of recognizing recursive script is not much more difficult than the problem of recognizing isolated characters. The performance of the recognition system appears to be better than that of any cursive script recognition system designed so far, even though extremely simplifying assumptions have been made in the stroke matching stage Thesaurus pattern recognition Other Terms online cursive script segmentation; feature extraction; stroke segmentation algorithm; curvature information processing; scale- space filtering; online author-dependent recognition system; handwritten data ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3485293 AbstractNos. B89071168; C89059609 References 4 U.S. Copyright Clearance Center Code CH2673-2/89/0000-1659$01.00 Country Pub. USA date 1162 ------------------------------------------------------------ Author Clippingdale, S.C.; Wilson, R.G.; Warwick Univ., Coventry, UK Title Least-squares image estimation on a multiresolution pyramid Source ICASSP-89: 1989 International Conference on Acoustics, Speech and Signal Processing (IEEE Cat. No.89CH2673-2); Part: Glasgow, UK; P art: 23-26 May 1989; Sponsored by: IEEE; New York, NY, USA; IEEE; 4 vol. 2833; 1989; pp. 1409-12 vol.3 Abstract A class of linear recursive image estimation methods, based on multiresolution image representations, is introduced. Although recursive, the estimators are causal not in the image plane but in a third dimension, that of the scale index. The estimators are efficient computationally and are in general suboptimal for a class of image models based explicitly on the scale-space underlying the multiresolution description. Methods are also presented for adapting the estimates to local image structure across a wide range of scales. The estimates are robust and outperform many of the techniques reported in the literature in terms of computational efficiency, signal-to-noise gain, and subjective appearance. A brief presentation of the theoretical basis of the methods is followed by experimental results and conclusions on the potential of the approach Thesaurus least squares approximations; picture processing Other Terms multiresolution pyramid; linear recursive image estimation methods; image plane; scale index; image models; computational efficiency; signal-to-noise gain ClassCodes B6140C; B0290F; C1250; C4130 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3485256 AbstractNos. B89071144; C89059595 References 16 U.S. Copyright Clearance Center Code CH2673-2/89/0000-1409$01.00 Country Pub. USA date 1162 ------------------------------------------------------------ Author Saint-Marc, P.; Chen, J.S.; Medioni, G.; Dept. of Electr. Eng. & Comput. Sci., Univ. of Southern California, Los Angeles, CA, USA Title Adaptive smoothing: a general tool for early vision Source Proceedings CVPR '89 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.89CH2752-4); Part: San Diego, CA, USA; Part: 4-8 June 1989; Sponsored by: IEEE; Washington, DC, USA; IEEE Comput. Soc. Press; xvii+693; 1989; pp. 618-24 Abstract The authors present a method to smooth a signal-whether it is an intensity image, a range image, or a contour-which preserves discontinuities and thus facilitates their detection. This is achieved by repeatedly convolving the signal with a very small averaging filter modulated by a measure of the signal discontinuity at each point. This process is related to the anisotropic diffusion reported by P. Perona and J. Malik (1987) but it has a much simpler formulation and is not subject to instability or divergence. Real examples show how this approach can be applied to the smoothing of various types of signals. The detected features do not move, and thus no tracking is needed. The last property makes it possible to derive a novel scale-space representation of a signal using a small number of scales. Finally, this process is easily implemented on parallel architectures: the running time on a 16 K connection machine is three orders of magnitude faster than on a serial machine Thesaurus computer vision Other Terms adaptive smoothing; computer vision; intensity image; range image; contour; signal discontinuity; scale-space representation; parallel architectures; 16 K connection machine ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3484911 AbstractNos. B89071136; C89059587 ISBN or SBN 0 8186 1952 x References 33 U.S. Copyright Clearance Center Code CH2752-4/89/0000-0618$01.00 Country Pub. USA date 1163 ------------------------------------------------------------ Author Mokhtarian, F.; Dept. of Comput. Sci., British Columbia Univ., Vancouver, BC, Canada Title Fingerprint theorems for curvature and torsion zero-crossings Source Proceedings CVPR '89 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.89CH2752-4); Part: San Diego, CA, USA; Part: 4-8 June 1989; Sponsored by: IEEE; Washington, DC, USA; IEEE Comput. Soc. Press; xvii+693; 1989; pp. 269-75 Abstract It has been shown by A.L. Yuille and T. Poggio (1983) that the scale-space image of a signal determines that signal uniquely up to constant scaling. Here, generalization of the proof given by Yuille and Poggio is presented. It is shown that the curvature scale-space image of a planar curvature determines the curvature uniquely, up to constant scaling and a rigid motion. The results show that a 1-D signal can be reconstructed using only one point from its scale-space image. This is an improvement of the result obtained by Yuille and Poggio Thesaurus pattern recognition; picture processing Other Terms fingerprint theorems; picture processing; pattern recognition; curvature; zero-crossings; scale-space image; Yuille; Poggio; constant scaling; rigid motion ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3484884 AbstractNos. B89071111; C89059555 ISBN or SBN 0 8186 1952 x References 10 U.S. Copyright Clearance Center Code CH2752-4/89/0000-0269$01.00 Country Pub. USA date 1163 ------------------------------------------------------------ Author Martens, J.B.O.S.; Majoor, G.M.M.; Inst. for Perception Res., Eindhoven, Netherlands Title The perceptual relevance of scale-space image coding Source Signal Processing; Signal Process. (Netherlands); vol.17, no.4; Aug. 1989; pp. 353-64 Abstract The authors summarize the so-called scale-space model and describe its application to image coding. In the model an image is passed through Gaussian filters of decreasing bandwidth. The variation between successively filtered responses is very systematic, so that little information is needed to pass between them. Starting from a low resolution version of the original image, they make a prediction for a higher resolution version. Only the prediction errors need be transmitted to recover this higher resolution picture. The process is repeated at a number of resolutions (called scales) in order to arrive at the original image. For data-reduction purposes, several approximations of these prediction errors can be studied. Evaluation of the resulting coded images is done by means of perceptual experiments. It is also shown that a one-to-one correspondence can be established between the different stages of the scale-space coder and a well-known model of the human visual system that is based on psychophysical data Thesaurus data compression; encoding; filtering and prediction theory; picture processing Other Terms perceptual relevance; image coding; scale-space model; Gaussian filters; prediction errors; higher resolution picture; data-reduction; human visual system; psychophysical data ClassCodes B6140C; B6120B; C1250 Article Type Theoretical / Mathematical; Experimental Coden SPRODR Language English RecordType Journal ControlNo. 3483291 AbstractNos. B89071090; C89059507 ISSN 01651684 References 21 U.S. Copyright Clearance Center Code 0165-1684/89/$3.50 Country Pub. Netherlands date 1165 ------------------------------------------------------------ Author Teh, C.-H.; Chin, R.T.; Dept. of Electr. & Comput. Eng., Wisconsin Univ., Madison, WI, USA Title On the detection of dominant points on digital curves Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.11, no.8; Aug. 1989; pp. 859-72 Abstract A parallel algorithm is presented for detecting dominant points on a digital closed curve. The procedure requires no input parameter and remains reliable even when features of multiple sizes are present on the digital curve. The procedure first determines the region of support for each point based on its local properties, then computes measures of relative significance (e.g. curvature) of each point, and finally detects dominant points by a process of nonmaximum suppression. This procedure leads to the observation that the performance of dominant points detection depends not only on the accuracy of the measure of significance, but also on the precise determination of the region of support. This solves the fundamental problem of scale factor selection encountered in various dominant point detection algorithms. The inherent nature of scale-space filtering in the procedure is addressed, and the performance of the procedure is compared to those of several other dominant point detection algorithms, using a number of examples Thesaurus computerised pattern recognition; computerised picture processing ; filtering and prediction theory; parallel algorithms Other Terms dominant point detection; computerised picture processing; computerised pattern recognition; digital curves; parallel algorithm; scale factor selection; scale-space filtering ClassCodes B6140C; C5260B; C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3477406 AbstractNos. B89071040; C89061512 ISSN 01628828 References 27 U.S. Copyright Clearance Center Code 0162-8828/89/0800-0859$01.00 Country Pub. USA date 1165 ------------------------------------------------------------ Author Gould, K.; Shah, M.; Dept. of Comput. Sci., Central Florida Univ., Orlando, FL, USA Title The trajectory primal sketch: a multi-scale scheme for representing motion characteristics Source Proceedings CVPR '89 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.89CH2752-4); Part: San Diego, CA, USA; Part: 4-8 June 1989; Sponsored by: IEEE; Washington, DC, USA; IEEE Comput. Soc. Press; xvii+693; 1989; pp. 79-85 Abstract Conventional approaches to dynamic scene analysis do not use motion itself explicitly for recognition. The authors propose a different approach for the use of motion in a computer vision system which uses the motion characteristics of moving objects without actually recovering the structure. In this approach, the extended trajectories followed by the objects are considered. It is argued that in many cases, where an object has a fixed and predefined motion, the trajectory of several points may serve to uniquely identify the object. In this approach, the trajectories are analyzed at multiple scales to identify important events corresponding to discontinuities in direction, speed, and acceleration using scale space. These important events are recorded in a presentation called trajectory primal sketch. Experimental results are presented graphically, demonstrating the potential value of this approach Thesaurus computer vision; computerised pattern recognition; computerised picture processing Other Terms feature extraction; computerised pattern recognition; trajectory primal sketch; motion characteristics; dynamic scene analysis; computer vision; multiple scales; scale space ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3471654 AbstractNos. B89063626; C89059532 ISBN or SBN 0 8186 1952 x References 6 U.S. Copyright Clearance Center Code CH2752-4/89/0000-0079$01.00 Country Pub. USA date 1163 ------------------------------------------------------------ Author Saund, E.; Dept. of Brain & Cognitive Sci., MIT, Cambridge, MA, USA Title Adding scale to the primal sketch Source Proceedings CVPR '89 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.89CH2752-4); Part: San Diego, CA, USA; Part: 4-8 June 1989; Sponsored by: IEEE; Washington, DC, USA; IEEE Comput. Soc. Press; xvii+693; 1989; pp. 70-8 Abstract The author shows how a scale dimension can be added to Marr's (1976) primal sketch to facilitate construction of multiscale descriptions of two-dimensional scales. In contrast to conventional scale-space approaches, this method omits any smoothing or blurring and performs grouping operations on symbolic shape tokens residing in a scale-space blackboard data structure. Two types of grouping operation are introduced: (1) fine-to-coarse aggregation of primitive-edge tokens builds coarser-scale edge maps from finer-scale information; and (2) pairwise grouping of symmetrically placed primitive edges gives rise to a primitive partial region token denoting curved-contour, primitive-corner, and bar events. The resulting collection of tokens makes the fundamental edge and region components of a shape's geometry available to later symbolic processes, leading to shape recognition or other tasks Thesaurus computerised pattern recognition; computerised picture processing ; data structures Other Terms Marr's primal sketch; 2D scales; feature extraction; computerised picture processing; pattern recognition; scale dimension; multiscale descriptions; grouping operations; symbolic shape tokens; scale-space blackboard data structure; fine-to-coarse aggregation; primitive-edge tokens; curved- contour; primitive-corner; bar events; shape recognition ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3471653 AbstractNos. B89063625; C89059531 ISBN or SBN 0 8186 1952 x References 25 U.S. Copyright Clearance Center Code CH2752-4/89/0000-0070$01.00 Country Pub. USA date 1163 ------------------------------------------------------------ Author Baker, H.H.; Artificial Intelligence Center, SRI Int., Menlo Park, CA, USA Title Building surfaces of evolution: the Weaving Wall Source International Journal of Computer Vision; Int. J. Comput. Vis. (Netherlands); vol.3, no.1; May 1989; pp. 51-71 Abstract Describes a three-dimensional surface-construction process designed for the analysis of image sequences. Named the Weaving Wall, the process operates over images as they arrive from a sensor, knitting together, along a parallel frontier, connected descriptions of images as they evolve over time. Although the Weaving Wall was developed to support a tracking mechanism for recovering the three-dimensional structure of a scene being traversed, other applications of the surface-building process have since become apparent. These include rendering and computation of tomographic medical data, display of higher- dimensional analytic functions, edge detection on the scale-space surface, and display and analysis of material fracture data. More generally, the Weaving Wall may be of use in representing the evolution of any two-dimensional imagery varying in a nearly continuous manner along a third dimension Thesaurus computer vision; computerised picture processing Other Terms Weaving Wall; three-dimensional surface-construction process; image sequences; descriptions of images; tracking mechanism; three-dimensional structure; surface-building process; tomographic medical data; edge detection; material fracture data ClassCodes C5260B Article Type Theoretical / Mathematical Coden IJCVEQ Language English RecordType Journal ControlNo. 3467785 AbstractNos. C89061520 ISSN 09205691 References 16 Country Pub. Netherlands date 1162 ------------------------------------------------------------ Author Chen, M.-H.; Yan, P.-F.; Dept. of Electr. Eng., State Univ. of New York, Stony Brook, NY, USA Title A multiscanning approach based on morphological filtering Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.11, no.7; July 1989; pp. 694-700 Abstract It is argued that the mathematical morphology method seems to be more reasonable and powerful in studying certain multiscaling vision problems than the approach that uses derivatives of Gaussian-shaped filters of different sizes. To show the validity of this method, the authors concentrated on an application that involves forming scale-space image of a 2-D shape using morphological opening filtering. A proof is given to show that morphological opening filtering has a property of not introducing additional zero-crossings as one moves to a coarser scale. This is a different result from the conclusion by A.L. Yuille and T.A. Poggio (ibid., vol.PAMI-8, Jan. 1986) that the Gaussian filter is the only filter with this property. In addition, opening filtering is computationaly simpler than the Gaussian filter Thesaurus filtering and prediction theory; picture processing Other Terms picture processing; image analysis; morphological filtering; multiscaling vision; scale-space image; 2-D shape; opening filtering ClassCodes B6140C; C1260; C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3466264 AbstractNos. B89063518; C89059802 ISSN 01628828 References 8 U.S. Copyright Clearance Center Code 0162-8828/89/0700-0694$01.00 Country Pub. USA date 1164 ------------------------------------------------------------ Author Zakharov, S.M.; Manykin, E.A. Title Optical image transformation by photon-echo signals in dynamic echo-holography Source Optika i Spektroskopiya; Opt. Spektrosk. (USSR); vol.65, no.2; Translated in: Optics and Spectroscopy; Translated in: Opt. Spectrosc. (USA); Translated in: vol.65, no.2; Translated in: Aug. 1988; pp. 249-51; Aug. 1988; pp. 419-23 Abstract Nonstationary optical image transformation by photon-echo signals is examined for usage of resonant media as dynamic spectral- selective holograms. The properties of scale space-time processing achieved in multilevel quantum systems are discussed Thesaurus holography; photon echo Other Terms photon-echo signals; dynamic echo-holography; optical image transformation; dynamic spectral-selective holograms; space- time processing; multilevel quantum systems ClassCodes A4240; A4265G; B4350; B4340 Article Type Theoretical / Mathematical Coden OSFMA3; OPSUA3 Language English RecordType Journal ControlNo. 3460120 AbstractNos. A89106787; B89062776 ISSN 00304034 ISSN (Trans) 0030400X References 18 U.S. Copyright Clearance Center Code 0030-400X/88/080249-03$05.00 Country Pub. USSR Country Pub. translation USA date 1152 ------------------------------------------------------------ Author Tsui, H.T.; Chu, K.C.; Dept. of Electron., Chinese Univ. of Hong Kong, Shatin, Hong Kong Title 3D object recognition by scale space feature tracking and subtemplate matching Source Intelligent Robots and Computer Vision; Part: Cambridge, MA, USA; Part: 7-11 Nov. 1988; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1002; 1989; pp. 609-16 Abstract A method to recognize 3D objects by detecting features at multiple scales and subtemplate matching is proposed. Depth map data of an object is first smoothed by Gaussian filtering at the coarsest scale and the Gaussian curvature at each point is computed. Extremal points are determined and an extremal point region (EPR) associated with each extremal point is defined. A spherical window which is invariant with rotation in 3-space is used to extract a surface patch around each extremal point for subtemplate matching. Processing and subtemplate matching are repeated at next finer scale to resolve ambiguities Thesaurus filtering and prediction theory; pattern recognition; picture processing Other Terms depth map data; 3D object recognition; picture processing; pattern recognition; scale space; feature tracking; subtemplate matching; Gaussian filtering; Gaussian curvature; extremal point region; spherical window ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3418937 AbstractNos. B89050108; C89045971 ISSN 0277786X References 17 Country Pub. USA date 1155 ------------------------------------------------------------ Author Rueff, M.; Fraunhofer Inst. for Manuf. Eng. & Autom., Stuttgart, West Germany Title Can scale space filtering enhance fractal analysis? Source Intelligent Robots and Computer Vision; Part: Cambridge, MA, USA; Part: 7-11 Nov. 1988; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.1002; 1989; pp. 136-43 Abstract Fractal dimensions are quantities which have been shown to be useful in the classification and segmentation of textures with scaling behaviour. Problems arising in the numerical determination of fractal dimensions are briefly mentioned. Scale space filtering techniques are suggested to overcome some of these problems which in particular are given with the detection of the limited scaling regions of natural textures Thesaurus computerised pattern recognition; computerised picture processing ; filtering and prediction theory; fractals Other Terms texture segmentation; computerised picture processing; computerised pattern recognition; scale space filtering; fractal analysis; scaling behaviour; fractal dimensions ClassCodes B6140C; C1250; C5260B; C1260; C6130B Article Type Theoretical / Mathematical Coden PSISDG Language English RecordType Conference ControlNo. 3418902 AbstractNos. B89050101; C89045950 ISSN 0277786X References 15 Country Pub. USA date 1155 ------------------------------------------------------------ Author Lu, Y.; Jain, R.C.; Dept. of Electr. Eng. & Comput. Sci., Michigan Univ., Ann Arbor, MI, USA Title Behavior of edges in scale space Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.11, no.4; April 1989; pp. 337-56 Abstract An analysis is presented of the behavior of edges in scale space for deriving rules useful in reasoning. This analysis of liner edges at different scales in images includes the mutual influence of edges and identifies at what scale neighboring edges start influencing the response of a Laplacian or Gaussian operator. Dislocation of edges, false edges, and merging of edges in the scale space are examined to formulate rules for reasoning in the scale space. The theorems, corollaries, and assertions presented can be used to recover edges, and related features, in complex images. The results reported include one lemma, three theorems, a number of corollaries and six assertions. The rigorous mathematical proofs for the theorems and corollaries are presented. These theorems and corollaries are further applied to more general situations, and the results are summarized in six assertions. A qualitative description as well as some experimental results are presented for each assertion Thesaurus artificial intelligence; computerised pattern recognition; computerised picture processing Other Terms edge behaviour; edge recovery; picture processing; computerised pattern recognition; scale space; reasoning; Laplacian or Gaussian operator; corollaries; assertions; complex images ClassCodes B6140C; C1250; C1230 Article Type Theoretical / Mathematical; Experimental Coden ITPIDJ Language English RecordType Journal ControlNo. 3401985 AbstractNos. B89044086; C89041490 ISSN 01628828 References 30 U.S. Copyright Clearance Center Code 0162-8828/89/0400-0337$01.00 Country Pub. USA date 1161 ------------------------------------------------------------ Author Piech, M.A.; Piech, K.R.; State Univ. of New York, Buffalo, NY, USA Title Hyperspectral interactions: invariance and scaling Source Applied Optics; Appl. Opt. (USA); vol.28, no.3; 1 Feb. 1989; pp. 481-9 Abstract The authors present new invariance and scaling results for scale space analysis of hyperspectral data. First, they note that a hyperspectral curve can be segmented into independent regions selected by features of scale space fingerprints. These fingerprint features are persistent inflection points that precisely locate major atmospheric features that define the regions. The strength and location of hyperspectral features in one atmospheric region are independent of features in other regions; as a result, hyperspectral analysis can be simplified to a region-by-region analysis. The authors then generate simple scaling and invariance rules for features within such a spectral region. They show that the scale of individual features is independent of the details of feature shape and depends only on the area of the feature. Interacting features in turn exhibit a fascinating bifurcation behavior: at large separations features behave independently; at smaller separations features interact and their scales are damped; below a critical separation distance (the bifurcation point) the features nest. The scales of features above the bifurcation point, the scales of the nested features, and the location of the bifurcation point depend only on the feature areas and not on shape-associated parameters of the individual features Thesaurus atmospheric spectra; remote sensing Other Terms remote sensing; invariance; scaling; scale space analysis; hyperspectral data; fingerprint features; inflection points; major atmospheric features; region-by-region analysis; bifurcation behavior; nested features; shape-associated parameters ClassCodes A9265H; B7730 Article Type Theoretical / Mathematical Coden APOPAI Language English RecordType Journal ControlNo. 3381244 AbstractNos. A89074019; B89040555 ISSN 00036935 References 15 U.S. Copyright Clearance Center Code 0003-6935/89/030481-09$02.00/0 Country Pub. USA date 1159 ------------------------------------------------------------ Author Chengsan Zhuang; Dept. of Comput. Sci. & Autom., Chengdu Univ. of Sci. & Technol., Sichuan, China Title Scale-based hierarchical description and matching of waveforms Source 9th International Conference on Pattern Recognition (IEEE Cat. No. 88CH2614-6); Part: Rome, Italy; Part: 14-17 Nov. 1988; Sponsored by: Int. Assoc. Pattern Recogition; Washington, DC, USA; IEEE Comput. Soc. Press; 2 vol. xxxvi+1299; 1988; pp. 1268-70 vol.2 Abstract An approach for obtaining hierarchical symbolical description of waveforms is proposed and a method for matching them is also given. The whole procedure is divided into three steps: first, scale-space filtering is applied to each waveform; second, peaks and valleys of all outputs of the filter are extracted; finally, a probability relaxation labeling algorithm is used to accomplish the matching. Results of experiments with synthetic data show that this approach is able to implement the rubberlike matching and the results obtained are not sensitive to noise Thesaurus pattern recognition Other Terms pattern recognition; scale based waveform description; hierarchical symbolical description; scale-space filtering; probability relaxation labeling algorithm; rubberlike matching ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3372736 AbstractNos. B89032585; C89030532 ISBN or SBN 0 8186 0878 1 References 8 U.S. Copyright Clearance Center Code CH2614-6/88/0000-1268$01.00 Country Pub. USA date 1155 ------------------------------------------------------------ Author Ranganathan, N.; Shah, M.; Dept. of Comput. Sci., Central Florida Univ., Orlando, FL, USA Title A scale-space chip Source 9th International Conference on Pattern Recognition (IEEE Cat. No. 88CH2614-6); Part: Rome, Italy; Part: 14-17 Nov. 1988; Sponsored by: Int. Assoc. Pattern Recogition; Washington, DC, USA; IEEE Comput. Soc. Press; 2 vol. xxxvi+1299; 1988; pp. 420-4 vol.1 Abstract Scale-space is a representation for detecting and organising intensity changes that occur at various scales in an image. A single-chip VLSI design is proposed for scale-space computation in one and two dimensions. The architecture of the chip is based on an algorithm that can provide speeds that are an order of magnitude higher than the speeds obtainable from the other systems proposed in the literature. The design uses the principles of modularity, expandability and parallelism, and fully utilizes the three properties of the Gaussian: symmetry, separability, and scaling. The proposed algorithm and the hardware architecture use a very high degree of pipelining and parallelism. The chip can be implemented in either nMOS or CMOS technology Thesaurus computer vision; computerised pattern recognition; computerised picture processing; integrated circuit technology; microprocessor chips; MOS integrated circuits; parallel architectures; pipeline processing; VLSI Other Terms computer vision; picture processing; chip architectures; image intensity change detection; intensity change organisation; nMOS technology; pattern recognition; scale-space chip; single-chip VLSI design; modularity; expandability; parallelism; symmetry; separability; pipelining; CMOS technology ClassCodes B2570D; B2570F; B1265F; C5260B; C1250; C5220; C5130 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 3363224 AbstractNos. B89030597; C89026356 ISBN or SBN 0 8186 0878 1 References 14 U.S. Copyright Clearance Center Code CH2614-6/88/0000-0420$01.00 Country Pub. USA date 1155 ------------------------------------------------------------ Author Huang Nan; Abbott, M.G.; Beattie, R.J.; Huazhong Univ. of Sci. & Technol., Wuhan, China Title Approaches to low level image processing for vision guided seam tracking systems Source 9th International Conference on Pattern Recognition (IEEE Cat. No. 88CH2614-6); Part: Rome, Italy; Part: 14-17 Nov. 1988; Sponsored by: Int. Assoc. Pattern Recogition; Washington, DC, USA; IEEE Comput. Soc. Press; 2 vol. xxxvi+1299; 1988; pp. 601-3 vol.1 Abstract Vision systems based on triangulation with active laser light sources are becoming widely used in robot arc welding. The sensor and preprocessing hardware provide a one-dimensional signal representing a cross section of the seam being tracked. This work describes and compares two different approaches to analyzing these signals as a precursor to matching them to predefined templates. One approach uses an expert systems methodology, while the other uses scale-space filtering Thesaurus arc welding; computer vision; expert systems; filtering and prediction theory; industrial robots; position control Other Terms industrial robots; position control; computer vision; low level image processing; vision guided seam tracking systems; triangulation; active laser light sources; robot arc welding; expert systems; scale-space filtering ClassCodes C3355F; C3120C; C3390; C5260B; C7410F; C1260 Article Type Practical Language English RecordType Conference ControlNo. 3351295 AbstractNos. C89025352 ISBN or SBN 0 8186 0878 1 References 2 U.S. Copyright Clearance Center Code CH2614-6/88/0000-0601$01.00 Country Pub. USA date 1155 ------------------------------------------------------------ Author Mokhtarian, F.; Dept. of Comput. Sci., British Columbia Univ., Vancouver, BC, Canada Title Evolution properties of space curves Source Second International Conference on Computer Vision (IEEE Cat. No. 88CH2664-1); Part: Tampa, FL, USA; Part: 5-8 Dec. 1988; Sponsored by: IEEE; Washington, DC, USA; IEEE Comput. Soc. Press; xiv+708; 1988; pp. 100-5 Abstract Several results are presented on the evolution properties of space curves which provide theoretical underpinning for the curvature and torsion scale space representation for space curves. It is shown that properties such as connectedness and closedness are preserved during evolution, that the centre of mass does not move as the curve evolves, that evolution is invariant under affine transformations of the curve such as uniform scaling, rotation, and translation, and that a space curve remains inside its convex hull during evolution. The two main theorems of this work show that there are strong constraints on the shape of a space curve in the neighborhood of a cusp point just before and just after the formation of that point Thesaurus computational geometry; computer vision; picture processing Other Terms space curves; curvature; torsion scale space representation; connectedness; closedness; centre of mass; affine transformations; uniform scaling; rotation; translation; convex hull; cusp point ClassCodes B6140C; B0250; B0290Z; C1250; C5260B; C1160; C4190 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 3350339 AbstractNos. B89025150; C89024169 ISBN or SBN 0 8186 0883 8 References 9 U.S. Copyright Clearance Center Code CH2664-1/88/0000-0100$01.00 Country Pub. USA date 1156 ------------------------------------------------------------ Author Perona, P.; Malik, J.; Dept. of Electr. Eng. & Comput. Sci., California Univ., Berkeley, CA, USA Title A network for multiscale image segmentation Source 1988 IEEE International Symposium on Circuits and Systems. Proceedings (Cat. No.88CH2458-8); Part: Espoo, Finland; Part: 7- 9 June 1988; Sponsored by: IEEE; New York, NY, USA; IEEE; 3 vol. 2915; 1988; pp. 2565-8 vol.3 Abstract Detecting edges of objects in their images is a basic problem in computational vision. The authors present the major ideas behind the use of scale space and anisotropic diffusion for edge detection, show that anisotropic diffusion can enhance edges, suggest a network implementation of anisotropic diffusion, and provide design criteria for obtaining networks performing scale space and edge detection. The results of a software implementation are shown Thesaurus computer vision; computerised pattern recognition; picture processing Other Terms picture processing; multiscale image segmentation; computational vision; scale space; anisotropic diffusion; edge detection; network implementation; design criteria; software implementation ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3339973 AbstractNos. B89025186; C89017673 References 12 U.S. Copyright Clearance Center Code CH2458-8/88/0000-2565$01.00 Country Pub. USA date 1150 ------------------------------------------------------------ Author Gardner, S.; US Naval Res. Lab., Washington, DC, USA Title Ultradiffusion, scale space transformation, and the morphology of neural networks Source IEEE International Conference on Neural Networks (IEEE Cat. No. 88CH2632-8); Part: San Diego, CA, USA; Part: 24-27 July 1988; Sponsored by: IEEE; New York, NY, USA; IEEE; 2 vol. (699+651); 1988; pp. 617-23 vol.1 Abstract The author proposes the scale-space transformation (SST) as a paradigm for information processing in biological neural networks. The SST concept includes scale-space, scale-time, and scale-space- time mappings. Hierarchical nonlinear (HNL) systems theory, together with the SST paradigm, causality requirements in the time domain, and uncertainty constraints in time and space domains, can be used to develop morphogenic models of biological neural networks. Since morphogenic models need only capture the functional modality of their physical counterparts, there may or may not be an observable resemblance to physical structure. To illustrate these concepts, the author discusses a morphogenic model of the mammalian visual system (MVS) in terms of SST mappings. As an example he uses an exponential retinotopic mapping, which is called the log Z SST (LZ SST). Using HNL and SST concepts, the author suggests a layered model of the MVS neural network Thesaurus neural nets; nonlinear systems; time-domain analysis Other Terms hierarchical nonlinear systems; morphology; scale-space transformation; biological neural networks; scale-space-time mappings; time domain; morphogenic model; mammalian visual system ClassCodes A8730E; A8710; C1290L Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3327891 AbstractNos. A89038193; C89017972 References 20 Country Pub. USA date 1151 ------------------------------------------------------------ Author Caelli, T.M.; Bischof, W.F.; Zhi-Qiang Liu; Dept. of Psychol., Queens Univ., Kingston, Ont., Canada Title Filter-based models for pattern classification Source Pattern Recognition; Pattern Recognit. (UK); vol.21, no.6; 1988; pp. 639-50 Abstract Considers a technique for pattern classification based upon the development of prototypes which capture the distinguishing features ('disjunctive prototypes') of each pattern class and, via cross-correlation with incoming test images, enable efficient pattern classification. The authors evaluate such a classification procedure with prototypes based on the images per se (direct code), Gabor scheme (multiple fixed filter representation) and an edge (scale space-based) coding scheme. The authors' analyses, the comparisons with human pattern classification performance, indicate that the edge-only disjunctive prototypes provide the most discriminating classification performance and are the more representative of human behaviour Thesaurus filtering and prediction theory; pattern recognition Other Terms pattern recognition; edge coding; pattern classification; disjunctive prototypes; cross-correlation; test images; Gabor scheme; edge-only disjunctive prototypes; human behaviour ClassCodes B6140C; C1250; C1260 Article Type Practical; Theoretical / Mathematical Coden PTNRA8 Language English RecordType Journal ControlNo. 3325466 AbstractNos. B89017974; C89017608 ISSN 00313203 References 13 U.S. Copyright Clearance Center Code 0031-3203/88/$3.00+.00 Country Pub. UK date 1145 ------------------------------------------------------------ Author Clark, J.J.; Div. of Appl. Sci., Harvard Univ., Cambridge, MA, USA Title Singularity theory and phantom edges in scale space Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.10, no.5; Sept. 1988; pp. 720-7 Abstract The process of detecting edges in a one-dimensional signal by finding the zeros of the second derivative of the signal can be interpreted as the process of detecting the critical points of a general class of contrast functions that are applied to the signal. It is shown that the second derivative of the contrast function at the critical point is related to the classification of the associated edge as being phantom or authentic. The contrast of authentic edges decreases with filter scale, while the contrast of phantom edges are shown to increase with scale. As the filter scale increases, an authentic edge must either turn into a phantom edge or join with a phantom edge and vanish. The points in the scale space at which these events occur are seen to be singular points of the contrast function. Using ideas from singularity, or catastrophy theory, the scale map contours near these singular points are found to be either vertical or parabolic Thesaurus catastrophe theory; filtering and prediction theory; pattern recognition; picture processing; signal processing Other Terms 1D signal; picture processing; edge detection; pattern recognition; singularity theory; phantom edges; scale space; contrast functions; filter scale; catastrophy theory; scale map contours ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 3310488 AbstractNos. B89017925; C89011043 ISSN 01628828 References 23 U.S. Copyright Clearance Center Code 0162-8828/88/0900-0720$01.00 Country Pub. USA date 1153 ------------------------------------------------------------ Author Sakou, H.; Yoda, H.; Ejiri, M.; Central Res. Lab., Hitachi Ltd., Tokyo, Japan Title An algorithm for matching distorted waveforms using a scale-based description Source Proceedings of IAPR Workshop on Computer Vision: Special Hardware and Industrial Applications; Part: Tokyo, Japan; Part: 12-14 Oct. 1988; Sponsored by: Int. Assoc. Pattern Recognition; Tokyo, Japan; Univ. Tokyo; x+459; 1988; pp. 329-34 Abstract Proposes a matching algorithm for two mutually-distorted waveforms each having partial differences in the hierarchical structure of its scale-space. This algorithm is applied to matching of two-dimensional shapes and to pattern-width measurement of semiconductor chip patterns obtained from a scanning electron microscope (SEM) Thesaurus computerised pattern recognition; integrated circuit technology; waveform analysis Other Terms scale-based description; matching algorithm; mutually-distorted waveforms; partial differences; scale-space; two-dimensional shapes; pattern-width measurement; semiconductor chip patterns; scanning electron microscope; SEM ClassCodes B6140C; B2570; C1250; C5260B Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 3306639 AbstractNos. B89010676; C89011133 References 7 Country Pub. Japan date 1154 ------------------------------------------------------------ Author Bertero, M.; Poggio, T.A.; Torre, V.; Dept. of Phys., Istituto Nazionale di Fisica Nucl., Genova, Italy Title Ill-posed problems in early vision Source Proceedings of the IEEE; Proc. IEEE (USA); vol.76, no.8; Aug. 1988; pp. 869-89 Abstract Mathematical results on ill-posed and ill-conditioned problems are reviewed and the formal aspects of regularization theory in the linear case are introduced. Specific topics in early vision and their regularization are then analyzed rigorously, characterizing existence, uniqueness, and stability of solutions. A fundamental difficulty that arises in almost every vision problem is scale, that is, the resolution at which to operate. Methods that have been proposed to deal with the problem include scale-space techniques that consider the behavior of the result across a continuum of scales. From the point of view of regulation theory, the concept of scale is related quite directly to the regularization parameter lambda . It suggested that methods used to obtained the optimal value of lambda may provide, either directly or after suitable modification, the optimal scale associated with the specific instance of certain problems Thesaurus computer vision Other Terms first state of processing; computer vision; ill posed conditions ; solution existence; solution uniqueness; early vision; ill- conditioned problems; regularization theory; analyzed rigorously ; existence; uniqueness; stability of solutions; vision problem; resolution; scale-space techniques; regulation theory; concept of scale; regularization parameter; optimal scale ClassCodes B6140C; C5260B; C1250 Article Type Bibliography/Literature Suvery; Theoretical / Mathematical Coden IEEPAD Language English RecordType Journal ControlNo. 3302796 AbstractNos. B89010577; C89013324 ISSN 00189219 References 89 U.S. Copyright Clearance Center Code 0018-9219/88/0800-0869$01.00 Country Pub. USA date 1152 ------------------------------------------------------------ Author Nguyen, D.T.; Ding-Yi Xu; Dept. of Electr. & Electron. Eng., Auckland Univ., New Zealand Title Scale based algorithm for recognition of blurred planar objects Source IEE Proceedings E (Computers and Digital Techniques); IEE Proc. E, Comput. Digit. Tech. (UK); vol.135, no.6; Nov. 1988; pp. 307-11 Abstract The paper presents an algorithm based on scale-space analysis, for the recognition of blurred planar objects. Apart from satisfying the usual requirements for invariance under translation, rotation and scaling, the algorithm is also invariant under blurring, that is, across all levels of detail or scales. The technique makes use of the spatial coincidence of the inflexion points on the object contour at all scales, and of the fact that no new such points are created as the object becomes more blurred. The algorithm therefore searches for the best match of these points at a single scale in the scale-space image. The algorithm was implemented on an IBM/AT in Modula-2 programming language, and was tested out on a group of 20 geographical maps of different sizes and at varying distances from the camera. A recognition rate of 95 to 100% and an average recognition time of 2.5 seconds were obtained by an efficient organisation of the template dictionary Thesaurus computerised pattern recognition; computerised picture processing Other Terms blurred planar objects recognition; scale based algorithm; scale-space analysis; translation; rotation; scaling; spatial coincidence; IBM/AT; Modula-2 programming language ClassCodes B6140C; C5260B; C5530 Article Type Practical Coden IPETD3 Language English RecordType Journal ControlNo. 3300476 AbstractNos. B89010540; C89013300 ISSN 01437062 References 7 U.S. Copyright Clearance Center Code 0143-7062/88/$3.00+0.00 Country Pub. UK date 1155 ------------------------------------------------------------ Author Meer, P.; Center for Autom. Res., Maryland Univ., College Park, MD, USA Title Simulation of constant size multiresolution representations on image pyramids Source Pattern Recognition Letters; Pattern Recognit. Lett. (Netherlands); vol.8, no.4; Nov. 1988; pp. 229-36 Abstract An image pyramid is a hierarchy of representations of the input derived by recursive smoothing and decimation. Image pyramids are built in log(image-size) time with the consecutive levels having their size and resolution reduced by a constant factor. Similar structures with the representations decreasing only in resolution but not in size are also of interest. The author simulates such constant size multiresolution representations of the input on image pyramids by increasing the number of values stored in the cells of the host structure. Constant size representations allow parallel processing in applications such as scale-space filtering and multiresolution edge detection Thesaurus computer vision; parallel processing Other Terms parallel processing; computer vision; constant size multiresolution representations; image pyramids; recursive smoothing; decimation; scale-space filtering; multiresolution edge detection ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical Coden PRLEDG Language English RecordType Journal ControlNo. 3299162 AbstractNos. B89010569; C89011077 ISSN 01678655 References 21 U.S. Copyright Clearance Center Code 0167-8655/88/$3.50 Country Pub. Netherlands date 1155 ------------------------------------------------------------ Author Koenderink, J.J.; Dept. of Eng. Sci., Oxford Univ., UK Title Scale-time Source Biological Cybernetics; Biol. Cybern. (West Germany); vol.58, no.3; 1988; pp. 159-62 Abstract Scale-space is a valuable tool in image processing and artificial vision. It is also of interest in the modeling of organic vision because mammalian visual systems appear to employ 'hardware' implementations of scale-space. Similar methods are used in the temporal domain, but such methods so far devised violate temporal causality. A conceptually simple and logically consistent scale- time which does not violate temporal causality, yet which conserves causality in the resolution domain at any given time is proposed. The filter kernels are not Gaussians (that would certainly lead to a violation of temporal causality) but are related to the Gaussians via a simple transformation of the time axis. They depend on a pair of parameters, one that has the character of a temporal delay and one that specifies the temporal resolution. In the limit for long delays (but fixed resolution) these kernels asymptotically approach the Gaussian again. Extensions of the theory towards a scale space-time are discussed Thesaurus computer vision; computerised picture processing; filtering and prediction theory; vision Other Terms computer vision; nonGaussian kernels; image processing; artificial vision; organic vision; mammalian visual systems; temporal causality; scale-time; resolution domain; filter kernels; temporal delay; temporal resolution ClassCodes A8710; A8732E; C1250; C1290L; C5260B; C1260 Article Type Theoretical / Mathematical Coden BICYAF Language English RecordType Journal ControlNo. 3279914 AbstractNos. A89009821; C89005736 ISSN 03401200 References 10 Country Pub. West Germany date 1145 ------------------------------------------------------------ Author Cooke, M.P.; Green, P.D.; Dept. of Comput. Sci., Sheffield Univ., UK Title On finding objects in spectrograms: a multiscale relaxation labelling approach Source Recent Advances in Speech Understanding and Dialog Systems. Proceedings of the NATO Advanced Institute; Part: Bad Windsheim, West Germany; Part: 5-18 July 1987; Sponsored by: NATO; Berlin, West Germany; Springer-Verlag; x+521; 1988; pp. 129-33 Editor Niemann, H.; Lang, M.; Sagerer, G. Abstract Describes a new technique for object finding in spectrograms, and illustrates the idea with an application to the formant-tracking task. Starting with a multiscale representation of speech spectra, a probabilistic relaxation labelling algorithm is applied to determine primitive interpretations of the spectral components. Finally, a cross-scale integration procedure enables the scale space to be collapsed in a principled manner. The techniques are illustrated with an example of voiced speech Thesaurus spectral analysis; speech analysis and processing Other Terms spectrograms; multiscale relaxation labelling approach; object finding; formant-tracking task; speech spectra; probabilistic relaxation labelling algorithm; primitive interpretations; spectral components; cross-scale integration procedure; scale space; voiced speech ClassCodes B6130 Article Type Practical Language English RecordType Conference ControlNo. 3278750 AbstractNos. B89003828 ISBN or SBN 3 540 19245 X References 4 Country Pub. West Germany date 1138 ------------------------------------------------------------ Author Mokhtarian, F.; Dept. of Comput. Sci., British Columbia Univ., Vancouver, BC, Canada Title Multi-scale description of space curves and three-dimensional objects Source Proceedings CVPR '88: The Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.88CH2605-4); Part: Ann Arbor, MI, USA; Part: 5-9 June 1988; Sponsored by: IEEE; Translated in: B19; Washington, DC, USA; IEEE Comput. Soc. Press; xv+975; 1988; pp. 298-303 Abstract The authors address the problem of representing the shape of three-dimensional or space curves. This problem is important since space curves can be used to model the shape of many three dimensional objects effectively and economically. A number of shape representation methods that operate on two-dimensional objects and can be extended to apply to space curves are reviewed briefly and their shortcomings discussed. Next, the concepts of curvature and torsion of a space curve are explained. Arc-length parametrization followed by Gaussian convolution is used to compute curvature and torsion on a space curve at varying levels of detail. Information of both the curvature and torsion of the curve over a continuum of scales are combined to produce the curvature and torsion scale-space images of the curve. These images are essentially invariant under rotation, uniform scaling, and translation of the curve and are used as a representation for it. The application of this technique to a common three- dimensional object is demonstrated. The proposed representation is then evaluated according to several criteria that any shape representation method should ideally satisfy Thesaurus computational geometry; pattern recognition Other Terms pattern recognition; arc-length parameterisation; shape representation; computational geometry; multiscale description; space curves; Gaussian ClassCodes B6140C; B0290Z; C1250; C4190 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3258529 AbstractNos. B88073322; C88060965 ISBN or SBN 0 8186 0862 5 References 32 U.S. Copyright Clearance Center Code CH2605-4/88/0000-0298$01.00 Country Pub. USA date 1150 ------------------------------------------------------------ Author Teh, C.-H.; Chin, R.T.; Dept. of Electr. & Comput. Eng., Wisconsin Univ., Madison, WI, USA Title A scale-independent dominant point detection algorithm Source Proceedings CVPR '88: The Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.88CH2605-4); Part: Ann Arbor, MI, USA; Part: 5-9 June 1988; Sponsored by: IEEE; Translated in: B08; Washington, DC, USA; IEEE Comput. Soc. Press; xv+975; 1988; pp. 229-34 Abstract A parallel algorithm for detecting dominant points on a digital closed curve is presented. The procedure requires no input parameter and remains reliable even when features of multiple sizes are present on the digital curve. The procedure first determines the region of support for each point based on its local properties, then computes measures of relative significance (e.g. curvature) of each point, and finally detects dominant points by a process of nonmaxima suppression. This procedure leads to an important observation that the performance of dominant points detection depends not only on the accuracy of the measure of significance, but mainly precise determination of the region of support. This solves the fundamental problem of scale factor selection encountered in various dominant point detection algorithms. The inherent nature of scale-space filtering in the procedure is addressed and the performance of the procedure is compared to those of several other dominant point-detection algorithms, using a number of examples Thesaurus computerised pattern recognition; parallel algorithms Other Terms computerised pattern recognition; scale-independent dominant point detection algorithm; scale factor selection ClassCodes B6140C; C5260B; C1250 Article Type Practical Language English RecordType Conference ControlNo. 3258520 AbstractNos. B88073313; C88063144 ISBN or SBN 0 8186 0862 5 References 13 U.S. Copyright Clearance Center Code CH2605-4/88/0000-0229$01.00 Country Pub. USA date 1150 ------------------------------------------------------------ Author Glass, J.R.; Zue, V.W.; Dept. of Electr. Eng. & Comput. Sci., MIT, Cambridge, MA, USA Title Multi-level acoustic segmentation of continuous speech Source ICASSP 88: 1988 International Conference on Acoustics, Speech, and Signal Processing (Cat. No.88CH2561-9); Part: New York, NY, USA; Part: 11-14 April 1988; Sponsored by: IEEE; Translated in: C07; New York, NY, USA; IEEE; 5 vol.2928; 1988; pp. 429-32 vol.1 Abstract As part of the goal to better understand the relationship between the speech signal and the underlying phonemic representation, the authors have developed a procedure that describes the acoustic structure of the signal. Acoustic events are embedded in a multi- level structure in which information ranging from coarse to fine is represented in an organized fashion. An analysis of the acoustic structure, using 500 utterances from 100 different talkers, show that it captures over 96% of the acoustic-phonetic events of interest with an insertion rate of less than 5%. The signal representation, and the algorithms for determining the acoustic segments and the multi-level structure are described. Performance results and a comparison with scale-space filtering is also included. Possible use of this segmental description for automatic speech recognition is discussed Thesaurus acoustic signal processing; speech analysis and processing; speech recognition Other Terms speech analysis; speech processing; multilevel acoustic segmentation; continuous speech recognition; speech signal; phonemic representation; acoustic; insertion rate; signal representation; scale-space filtering; automatic speech recognition ClassCodes A4370; A4360; B6130; B6140; C1250C Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 3257059 AbstractNos. A88128461; B88072795; C88061042 References 8 U.S. Copyright Clearance Center Code CH2561-9/88/0000-0429$1.00 Country Pub. USA date 1148 ------------------------------------------------------------ Author Mackworth, S.K.; Mokhtarian, F.; Dept. of Comput. Sci., British Columbia Univ., Vancouver, BC, Canada Title The renormalized curvature scale space and the evolution properties of planar curves Source Proceedings CVPR '88: The Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.88CH2605-4); Part: Ann Arbor, MI, USA; Part: 5-9 June 1988; Sponsored by: IEEE; Translated in: B22; Washington, DC, USA; IEEE Comput. Soc. Press; xv+975; 1988; pp. 318-26 Abstract The curvature scale-space image of a planar curve is computed by convolving a path-based parametric representation of the curve with a Gaussian function of variance sigma /sup 2/, extracting the zeros of curvature of the convolved curves and combining them in a scale space representation of the curve. For any given curve Gamma , the process of generating the ordered sequence of curves ( Gamma /sub sigma / mod sigma >or=0) is the evolution of Gamma . It is shown that the normalized arc length parameter of a curve is, in general, not the normalized arch length parameter of a convolved version of that curve. A novel method of computing the curvature scale space image reparametrizes each convolved curve by its normalized arc length parameter. Zeros of curvature are then expressed in that new parametrization. The result is the renormalized curvature scale-space image and is more suitable for matching curves similar in shape. Scaling properties of planar curves and the curvature scale space image are also investigated. It is shown that no new curvature zero-crossings are created at the higher scales of the curvature scale space image of a planar curve in C/sub 1/ if the curve remains in C/sub 1/ during evolution. Several results are presented on the preservation of various properties of planar curves under the evolution process Thesaurus computational geometry Other Terms computational geometry; evolution properties; curvature scale- space image; planar curve; path-based parametric representation; Gaussian function ClassCodes C4190 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3248007 AbstractNos. C88062342 ISBN or SBN 0 8186 0862 5 References 10 U.S. Copyright Clearance Center Code CH2605-4/88/0000-0318$01.00 Country Pub. USA date 1150 ------------------------------------------------------------ Author Ranganathan, N.; Shah, M.; Dept. of Comp. Sci., Central Florida Univ., Orlando, FL, USA Title A VLSI architecture for computing scale space Source Computer Vision, Graphics, and Image Processing; Comput. Vis. Graph. Image Process. (USA); vol.43, no.2; Translated in: A04; Aug. 1988; pp. 178-204 Abstract Meaningful information about a scene is captured in the intensity changes in an image. These intensity changes occur at various scales depending on their physical origins. Scale-space generated by applying the Laplacian of Gaussian edge detector to the image at a continuum of scales is a powerful representation for detecting and organizing these intensity changes symbolically and has proved to be very useful for one-dimensional signals. The high computational cost of generating scale-space in two dimensions has restricted its use in images. This paper proposes a very efficient single chip VLSI design for scale-space computation in one and two dimensions. The architecture of the chip is based on an algorithm that can provide speeds that are of an order to magnitude higher than the speeds obtainable from other systems proposed in the literature. The design uses the principles of modularity, expandability, and parallelism, and fully utilizes the three properties of Gaussian symmetry, separability, and scaling. Moreover, the proposed algorithm does not approximate the Laplacian of the Gaussian operator; it uses instead four one-dimensional convolutions to obtain the computations in two dimensions. The proposed architecture has not been built Thesaurus computer vision; computerised pattern recognition; computerised picture processing; microprocessor chips; VLSI Other Terms computer vision; picture processing; VLSI architecture; intensity changes; Gaussian edge detector; scale-space computation; modularity; expandability; parallelism; one- dimensional convolutions ClassCodes B2570; C5130; C5260B Article Type Practical Coden CVGPDB Language English RecordType Journal ControlNo. 3241831 AbstractNos. B88064637; C88063071 ISSN 0734189X References 32 U.S. Copyright Clearance Center Code 0734-189X/88/$3.00 Country Pub. USA date 1152 ------------------------------------------------------------ Author Kass, M.; Witkin, A.; Terzopoulos, D.; Schlumberger Palo Alto Res., CA, USA Title Snakes: active contour models Source International Journal of Computer Vision; Int. J. Comput. Vis. (Netherlands); vol.1, no.4; Translated in: A03; 1987; pp. 321-31 Abstract A snake is an energy-minimizing spline guided by external constraint forces and influenced by image forces that pull it toward features such as lines and edges. Snakes are active contour models: they lock onto nearby edges, localizing them accurately. Scale-space continuation can be used to enlarge the capture region surrounding a feature. Snakes provide a unified account of a number of visual problems, including detection of edges, lines, and subjective contours; motion tracking; and stereo matching Thesaurus computer vision Other Terms scale-space continuation; active contour models; snake; energy- minimizing spline; external constraint forces; image forces; lines; edges; capture region; subjective contours; motion tracking; stereo matching ClassCodes B6140C; C1250; C5260B Article Type Practical; Theoretical / Mathematical Language English RecordType Journal ControlNo. 3235173 AbstractNos. B88066638; C88057363 ISSN 09205691 References 25 Country Pub. Netherlands date 1132 ------------------------------------------------------------ Author Piech, M.A.; Dept. of Math., State Univ. of New York, Buffalo, NY, USA Title Comments on fingerprints of two-dimensional edge models Source Computer Vision, Graphics, and Image Processing; Comput. Vis. Graph. Image Process. (USA); vol.42, no.3; Translated in: A07; June 1988; pp. 381-6 Abstract M. Shah, A. Sood and R. Jain have published an interesting scale- space analysis of pulse and staircase edge models, both one- and two-dimensional (see ibid., vol.34, p.321-43, 1986). This note comments upon Shah, Sood and Jain's analysis of the two- dimensional step edge, pulse edge and staircase edge models. Their derivation of the scale-space images and fingerprints can be simplified by taking advantage of a key geometric feature of Gaussian filters, namely rotational invariance. The fingerprints of these three models can easily and directly be deduced geometrically from the fingerprints of the one-dimensional models. The fingerprints should be viewed as cylinders over a base curve which is precisely the fingerprint of the corresponding one- dimensional edge model. In this way fingerprints of the two- dimensional models can be immediately visualized from their one- dimensional counterparts. The authors also demonstrate that the range of influence of one edge upon another edge located a distance d away begins at a scale of d/3 Thesaurus filtering and prediction theory; pattern recognition; picture processing Other Terms pulse edge models; fingerprints; two-dimensional edge models; scale-space analysis; staircase edge models; Gaussian filters; rotational invariance ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden CVGPDB Language English RecordType Journal ControlNo. 3170388 AbstractNos. B88046158; C88037751 ISSN 0734189X References 2 U.S. Copyright Clearance Center Code 0734-189X/88/$3.00 Country Pub. USA date 1150 ------------------------------------------------------------ Author Fan, T.J.; Medioni, G.; Nevatia, R.; Dept. of Electr. Eng., Univ. of Southern California, Los Angeles, CA, USA Title 3-D surface description using curvature properties Source Optical and Digital Pattern Recognition; Part: Los Angeles, CA, USA; Part: 13-15 Jan. 1987; Sponsored by: SPIE; Proceedings of the SPIE - The International Society for Optical Engineering; vol.754; Translated in: A15; 1987; pp. 100-6 Abstract Presents a method to extract and represent significant physical properties of a surface, using curvature properties of this surface. The authors compute curvature in 4 different directions and detect extrema and zero-crossings for each of these one dimensional curves. Since this computation is very noise sensitive, they filter these features using a scale-space tracking approach: they smooth the image with Gaussian masks of increasing variance, detecting features at the smoothest level and localizing them at the original one. They then partially group these features into junctions which correspond to significant physical properties, such as depth discontinuities, surface discontinuities, smooth extrema, and link them into curves. They believe that these descriptions capture most of the information present in the original image, but are more suited to further processing, such as matching with a model. They then illustrate this technique with several examples Thesaurus pattern recognition Other Terms surface description; curvature properties; extrema; zero- crossings; scale-space tracking approach; Gaussian masks ClassCodes B6140C; C1250; C5260B Article Type Theoretical / Mathematical; Experimental Coden PSISDG Language English RecordType Conference ControlNo. 3169681 AbstractNos. B88046201; C88037777 ISSN 0277786X References 25 Country Pub. USA date 1132 ------------------------------------------------------------ Author Bischof, W.F.; Caelli, T.; Dept. of Psychol., Alberta Univ., Edmonton, Alta., Canada Title Parsing scale-space and spatial stability analysis Source Computer Vision, Graphics, and Image Processing; Comput. Vis. Graph. Image Process. (USA); vol.42, no.2; Translated in: A03; May 1988; pp. 192-205 Abstract The scale-space S(x, sigma ) of a signal I(x) is defined as the space of the zero-crossings from ( Del /sup 2/G( sigma )*I(x)), where G is a Gaussian filter. The authors present a new method for parsing scale-space, spatial stability analysis, that allows the localization of region boundaries from scale space. Spatial stability analysis is based on the observation that zero- crossings of region boundaries remain spatially stable over changes in filter scale. It is shown that spatial stability analysis leads to an edge detection scheme with good noise resilience characteristics and that it can lead to improvements in 'shape from texture' methods Thesaurus computerised pattern recognition; computerised signal processing; filtering and prediction theory Other Terms computerised pattern recognition; zero-crossings; Gaussian filter; parsing; spatial stability analysis; scale space; edge detection; noise resilience ClassCodes C1250; C1260 Article Type Theoretical / Mathematical Coden CVGPDB Language English RecordType Journal ControlNo. 3160349 AbstractNos. C88037744 ISSN 0734189X References 29 U.S. Copyright Clearance Center Code 0734-189X/88/$3.00 Country Pub. USA date 1149 ------------------------------------------------------------ Author Perona, P.; Malik, J.; Dept. of Electr. Eng. & Comput. Sci., California Univ., Berkeley, CA, USA Title Scale space and edge detection using anisotropic diffusion Source Proceedings of the IEEE Computer Society Workshop on Computer Vision (Cat. No.87TH0210-5); Part: Miami Beach, FL, USA; Part: 30 Nov.-2 Dec. 1987; Translated in: A03; Washington, DC, USA; IEEE Comput. Soc. Press; xi+370; 1987; pp. 16-22 Abstract The scale-space technique introduced by A.P. Witkin (1983) involves generating coarser resolution images by convolving the original image with a Gaussian kernel, or equivalently by using the original image as the initial condition of a diffusion process. This approach has a major drawback; it is difficult to obtain accurately the locations of the 'semantically meaningful' edges at coarse scales. The authors suggest a novel definition of scale-space and introduce a class of algorithms that realize it using anisotropic diffusion. The diffusion coefficient is chosen to vary spatially in such a way as to encourage intraregion smoothing in preference to interregion smoothing. It is shown that the 'no new maxima should be generated at coarse scales' property of conventional scale space is preserved. As the region boundaries in the proposed approach remain sharp, a high quality edge detector which successfully utilizes global information is obtained. Experimental results are shown on a couple of images. The algorithm involves simple, local operations replicated over the image, making parallel hardware implementation feasible Thesaurus diffusion; pattern recognition Other Terms pattern recognition; edge detection; anisotropic diffusion; scale-space technique; Gaussian kernel; intraregion smoothing; global information ClassCodes B0240Z; B6140C; C1140Z; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3118766 AbstractNos. B88028092; C88022065 ISBN or SBN 0 8186 4779 5 References 13 U.S. Copyright Clearance Center Code TH0210-5/87/0000-0016$01.0 Country Pub. USA date 1142 ------------------------------------------------------------ Author Sato, M.; Wada, T.; Res. Lab. of Precision Machinery & Electron., Tokyo Inst. of Technol., Yokohama, Japan Title A hierarchical representation of generalized waveforms Source Transactions of the Institute of Electronics, Information and Communication Engineers D; Trans. Inst. Electron. Inf. Commun. Eng. D (Japan); vol.J70D, no.11; Translated in: A20; Nov. 1987; pp. 2154-9 Abstract A method to generate a tree is proposed, which represents the hierarchical structure of waveforms by scale space filtering. The generated tree called the structure line describes the relation of the convex and the concave regions of scale space filtered waveform called the generalized waveform. The structure line is defined by some derivatives of the generalized waveform, and holds the same structure as a trinary tree topologically. The properties of structure line are shown, and also that this method is effective to represent the waveform hierarchically Thesaurus computerised pattern recognition; hierarchical systems Other Terms scale space filtering; hierarchical representation; generalized waveforms; hierarchical structure of waveforms; scale space filtering ClassCodes C1250 Article Type Theoretical / Mathematical Coden DJTDE2 Language Japanese RecordType Journal ControlNo. 3087813 AbstractNos. C88016244 ISSN 0374468X References 7 Country Pub. Japan date 1142 ------------------------------------------------------------ Author Nguyen, D.T.; Xu, D.-Y.; Dept. of Electr. & Electron. Eng., Auckland Univ., New Zealand Title Efficient technique for scale-space imaging of planar objects Source Electronics Letters; Electron. Lett. (UK); vol.23, no.24; Translated in: C08; 19 Nov. 1987; pp. 1326-7 Abstract The scale-space image, i.e. the plot of the location of the inflection points against the scale of the Gaussian smoothing filter, is an effective description for planar object recognition. Recent techniques of scale-space imaging involve four convolutions for each scale, while the authors' technique by smoothing the curvature of the object contour directly requires only one convolution per scale Thesaurus computerised picture processing Other Terms inflection points location; curvature smoothing; scale-space imaging of planar objects; scale-space image; Gaussian smoothing filter; planar object recognition; one convolution per scale ClassCodes B6140C; C1250; C1260 Article Type Theoretical / Mathematical Coden ELLEAK Language English RecordType Journal ControlNo. 3070945 AbstractNos. B88014758; C88010536 ISSN 00135194 References 4 U.S. Copyright Clearance Center Code 0013-5194/87/$2.00+0.00 Country Pub. UK date 1142 ------------------------------------------------------------ Author Shimizu, E.; Matsushita, K.; Takahashi, H. Title Sensors and new signal processing techniques Source Systems and Control; Syst. Control (Japan); vol.31, no.2; Translated in: T01; Feb. 1987; pp. 87-94 Abstract Sensors and signal processing techniques have been developed to meet application requirements. This paper explains new signal processing techniques in three classifications: orthogonal- transfer, matrix, and new signal processing techniques. The orthogonal-transfer techniques include the Fourier transfer, the sampling Fourier transfer, the fast Fourier transfer, the Adamahl transfer, and the Fourier descripter. The matrix technique explanation focuses on the coordinate transfer type computing for image processing. As the new signal processing technique, the optical flow, maximum entropy spectrum, and scale space filtering are introduced Thesaurus filtering and prediction theory; Fourier transforms; image sensors; picture processing; signal processing Other Terms picture processing; signal processing techniques; orthogonal- transfer; sampling Fourier transfer; fast Fourier transfer; Adamahl transfer; Fourier descripter; matrix; coordinate transfer; image processing; optical flow; maximum entropy spectrum; scale space filtering ClassCodes A4230V; B0230; B6140; B7230; B7230G; C1130; C1250; C1260 Article Type General or Review; Theoretical / Mathematical Coden SYCNA9 Language Japanese RecordType Journal ControlNo. 3057091 AbstractNos. A88012591; B88008778; C88005650 ISSN 03744507 References 18 Country Pub. Japan date 1133 ------------------------------------------------------------ Author Huttenlocher, D.P.; Ullman, S.; Artificial Intelligence Lab., MIT, Cambridge, MA, USA Title Object recognition using alignment Source Image Understanding Workshop Proceedings; Part: Los Angeles, CA, USA; Part: 23-25 Feb. 1987; Sponsored by: Defense Adv. Res. Projects Agency; Translated in: C01; Los Altos, CA, USA; Morgan Kaufmann; 2 vol. vi+1000; 1987; pp. 370-80 vol.1 Abstract The paper presents an approach to recognition whereby an object is first aligned with an image using a small number of pairs of model and image features, and then the aligned model is compared directly against the image. For instance, the position, orientation, and scale of an object in three-space can be determined from three pairs of corresponding model and image features. By using a small fixed number of features to determine position and orientation, the alignment process avoids structuring the recognition problem as an exponential search. To demonstrate the method, some examples of the recognition of flat rigid objects with arbitrary three-dimensional position, orientation, and scale, from a single two-dimensional image, are given. The recognition system chooses features for alignment using a scale-space segmentation of edge contours. Finally, the method is extended to the domain of rigid objects in general Thesaurus pattern recognition Other Terms object recognition; alignment; position; orientation; scale; scale-space segmentation; edge contours ClassCodes C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3049253 AbstractNos. C88005684 ISBN or SBN 0 934613 36 2 References 19 Country Pub. USA date 1133 ------------------------------------------------------------ Author Lenz, R.; Linkoping Univ., Sweden Title Rotation-invariant operators and scale-space filtering Source Pattern Recognition Letters; Pattern Recognit. Lett. (Netherlands); vol.6, no.3; Translated in: A01; Aug. 1987; pp. 151-4 Abstract Analysis of images at different spatial scales is known to be an important tool in the processing of images. One of the most popular procedures in this branch of image processing is based on the zero-crossings of the Laplacian. The author analyses the 'Laplacian of the Gaussian' approach with the help of rotation- invariant operators and shows how the original image is related to the Gaussian (and Laplace-) filtered version Thesaurus filtering and prediction theory; Laplace transforms; picture processing Other Terms picture processing; Gaussian filtering; Laplacian filtering; scale-space filtering; image processing; rotation-invariant operators ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden PRLEDG Language English RecordType Journal ControlNo. 3031460 AbstractNos. B88003189; C88000560 ISSN 01678655 References 7 U.S. Copyright Clearance Center Code 0167-8655/87/$3.50 Country Pub. Netherlands date 1139 ------------------------------------------------------------ Author Witkin, A.; Terzopoulos, D.; Kass, M.; Schlumberger Palo Alto Res., CA, USA Title Signal matching through scale space Source International Journal of Computer Vision; Int. J. Comput. Vis. (Netherlands); vol.1, no.2; Translated in: A02; 1987; pp. 133-44 Abstract Given a collection of similar signals that have been deformed with respect to each other, the general signal-matching problem is to recover the deformation. The authors formulate the problem as the minimization of an energy measure that combines a smoothness term and a similarity term. The minimization reduces to a dynamic system governed by a set of coupled, first-order, differential equations. The dynamic system finds an optimal solution at a coarse scale and then tracks it continuously to a fine scale. Among the major themes in recent work on visual signal matching have been the notions of matching as constrained optimization, of variational surface reconstruction and of coarse- to-fine matching. The authors' solution captures these in a precise, succinct and unified form. Results are presented for one- dimensional signals, a motion sequence and a stereo pair Thesaurus computer vision; minimisation; pattern recognition; picture processing Other Terms energy-measure minimisation; coupled first-order differential equations; computer vision; scale space; visual signal matching ; constrained optimization; variational surface reconstruction; coarse-to-fine matching; one-dimensional signals; motion sequence; stereo pair ClassCodes C1180; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Journal ControlNo. 3026029 AbstractNos. C88000553 ISSN 09205691 References 22 Country Pub. Netherlands date 1132 ------------------------------------------------------------ Author Piech, M.A.; Piech, K.R.; Dept. of Math., State Univ. of New York, Buffalo, NY, USA Title Symbolic representation of hyperspectral data Source Applied Optics; Appl. Opt. (USA); vol.26, no.18; Translated in: D09; 15 Sept. 1987; pp. 4018-26 Abstract The authors have developed a symbolic representation of hyperspectral data using the scale space techniques of Witkin (1983). They created a scale space image of hyperspectral data from convolution with Gaussian masks and then a fingerprint that extracts individual features from the original data. The fingerprint provides a context that pairs inflection points and assigns them to a feature, generates a measure of importance for each feature, and relates features to each other. The representation is an ordered sequence of triplets containing a measure of importance related to the area of each feature and the left and right inflection points of the feature. The description is compact, quantitative, and hierarchical, describing the hyperspectral curve by its most important structural features first, followed by features of lesser importance Thesaurus remote sensing; spectroscopy Other Terms symbolic representation; hyperspectral data; scale space techniques; scale space image; convolution; Gaussian masks ClassCodes A0650D; A0765 Article Type Practical; Theoretical / Mathematical Coden APOPAI Language English RecordType Journal ControlNo. 3021291 AbstractNos. A88000353 ISSN 00036935 References 10 U.S. Copyright Clearance Center Code 0003-6935/87/184018-09/$02.00/0 Country Pub. USA date 1140 ------------------------------------------------------------ Author Cyganski, D.; Orr, J.A.; Cott, T.A.; Dodson, R.; Dept. of Electr. Eng., Worcester Polytech. Inst., MA, USA Title Implementation of a tensor differential scale space system Source Proceedings of the Nineteenth Southeastern Symposium on System Theory (Cat. No.TH0180-0); Part: Clemson, SC, USA; Part: 15-17 March 1987; Sponsored by: IEEE; Translated in: A04; Washington, DC, USA; IEEE Comput. Soc. Press; xvii+575; 1987; pp. 18-22 Abstract A variation on the scale-space representation of planar curves is introduced, using a parameterization and curvature definition which extends the range of cases for which scale-space methods are useful. Previously, the scale-space approach to image registration and identification was unable to deal with images skewed by a change in angle between object plane and camera line of sight. The new approach, using tensor curve differentials to assemble affine invariant measures, eliminates this restriction, and is applicable to image pairs related by any affine transformation. Numerical methods are described for the implementation of such a system, and the robustness of this implementation for image distortion due to spatial quantization is illustrated. By using local averaging in the form of least- squares low-order polynomial fits, derivatives of the first through third order of sufficient quality for affine scale-space representation can be obtained Thesaurus pattern recognition; picture processing; tensors Other Terms picture processing; pattern recognition; image identification; skewed images; tensor differential scale space system; image registration; image distortion; spatial quantization; local averaging; least-squares low-order polynomial fits; affine scale-space representation ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 3000815 AbstractNos. B87069373; C87057279 ISBN or SBN 0 8186 0717 3 References 8 U.S. Copyright Clearance Center Code 0094-2898/87/0000-0018$01.00 Country Pub. USA date 1134 ------------------------------------------------------------ Author Meer, P.; Baugher, S.; Rosenfeld, A.; Center for Autom. Res., Maryland Univ., College Park, MD, USA Title Segmentation of multiscale curves by chain pyramids Source Topical Meeting on Machine Vision. Technical Digest Series Vol.12 (papers in summary form only received); Part: Incline Village, NV, USA; Part: 18-20 March 1987; Sponsored by: Opt. Soc. America; Translated in: B20; Washington, DC, USA; Opt. Soc. America; ix+180; 1987; pp. 176-9 Abstract A planar curve may convey information at several levels of detail. Locating the curve extrema with the best precision, while trivial for human perception, presents considerable difficulties for computer vision. Inconsistent local data must first be eliminated through soothing. The extrema are then found on the smoothed curve and projected onto the input. Smoothing in the Fourier domain is convenient but mapping extrema through the different levels of resolution is difficult. The scale space approach may overcome some of the difficulties but is computationally expensive. The authors propose a new method which performs smoothing in the image domain. The processing is parallel and the extrema mapping is immediate. The method is a particular case of a new approach to processing planar curve data Thesaurus computerised pattern recognition; computerised picture processing Other Terms computerised picture processing; segmentation; computerised pattern recognition; multiscale curves; chain pyramids; planar curve; computer vision; smoothing; image domain; extrema mapping ClassCodes C1250; C5260 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 2988226 AbstractNos. C87059740 ISBN or SBN 0 936659 47 5 References 6 Country Pub. USA date 1134 ------------------------------------------------------------ Author Huttenlocher, D.P.; Ullman, S.; Artificial Intelligence Lab., MIT, Cambridge, MA, USA Title Object recognition using alignment Source Proceedings of the First International Conference on Computer Vision (Cat. No.87CH2465-3); Part: London, UK; Part: 8-11 June 1987; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: A11; Washington, DC, USA; IEEE Comput. Soc. Press; xii+734; 1987; pp. 102-11 Abstract An approach to recognition is presented in which an object is first aligned with an image using a small number of pairs of model and image features, and then the aligned model is compared directly against the image. For instance, the position, orientation, and scale of an object in three-space can be determined from three pairs of corresponding model and image points. By using a small field number of features to determine position and orientation, the alignment method avoids structuring the recognition process as an exponential search. To demonstrate the method, some examples are presented of recognizing flat rigid objects with arbitrary three-dimensional position, orientation, and scale, from a single two-dimensional image. The recognition system chooses features for alignment using a scale-space segmentation of edge contours, which yields relatively distinctive feature labels. The method is extended to the domain of nonflat objects as well Thesaurus pattern recognition Other Terms object recognition; computer vision; alignment; image; image features; position; orientation; arbitrary three-dimensional position; scale; scale-space segmentation; edge contours ClassCodes C1250 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 2987827 AbstractNos. C87057244 ISBN or SBN 0 8186 0777 7 References 15 U.S. Copyright Clearance Center Code CH2465-3/87/0000-0102$01.00 Country Pub. USA date 1137 ------------------------------------------------------------ Author Mallat, S.G.; Dept. of Comput. & Inf. Sci., Pennsylvania Univ., Philadelphia, PA, USA Title Scale change versus scale space representation Source Proceedings of the First International Conference on Computer Vision (Cat. No.87CH2465-3); Part: London, UK; Part: 8-11 June 1987; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: C25; Washington, DC, USA; IEEE Comput. Soc. Press; xii+734; 1987; pp. 592-6 Abstract It is acknowledged in the computer vision literature that a multiscale decomposition of images provides a useful representation for many image processing algorithms. It is shown that the wavelet theory, recently developed by Y. Meyer (see Bourbaki seminar, no.662, 1985/6) enables the author to define precisely the concept of scaling transformation. For a multiscale analysis of images, he does not want to process the signal at each scale because the information is redundant. After processing the signal at a scale s/sub 0/, it is more efficient to analyze only the additional details which are available at a higher scale (greater resolution) s/sub 1/. The wavelet theory enables the author to mathematically define this difference of information, and it is pointed out that it can be efficiently computed with a pyramid transform. This leads to a decomposition of the image in a set of frequency channels with an orientation selectivity which is called the scale-change representation. This representation is particularly well adapted to computer vision tasks such as texture analysis, edge detection and matching algorithms Thesaurus computer vision; parallel algorithms; pattern recognition; picture processing Other Terms multiscale image decomposition; Meyer wavelet theory; redundant information; parallel algorithm; scale space representation; computer vision; image processing; scaling transformation; pyramid transform; scale-change representation; texture analysis ; edge detection; matching ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 2977760 AbstractNos. B87062369; C87051427 ISBN or SBN 0 8186 0777 7 References 5 U.S. Copyright Clearance Center Code CH2465-3/87/0000-0592$01.00 Country Pub. USA date 1137 ------------------------------------------------------------ Author Hummel, R.; Moniot, R.; New York Univ., NY, USA Title Solving ill-conditioned problems by minimizing equation error Source Proceedings of the First International Conference on Computer Vision (Cat. No.87CH2465-3); Part: London, UK; Part: 8-11 June 1987; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: C15; Washington, DC, USA; IEEE Comput. Soc. Press; xii+734; 1987; pp. 527-33 Abstract Ill-conditioned problems arise frequently in computer vision because the image information contains noise and ambiguities such that the true identity of the scene is not uniquely specified. A method in numerical analysis is considered for solving 'inverse problems' when the degraded scene has undergone a sequence of steps modeled by a known equation. Reconstruction can then be attempted by finding a solution that minimizes the deviation from that equation. The method is exemplified by its application to the deblurring problem. In this problem, deblurring is achieved by computing a succession of images, each slightly deblurred from the previous, such that the complete set satisfies the equations specifying the diffusion process of blurring. The method can be viewed as an approach to regularization for problems in which a scale-space parameter can be used to separate the information extracted from the image Thesaurus computer vision; error analysis; numerical methods; picture processing Other Terms equation error minimisation; degraded images; image reconstruction; ill-conditioned problems; computer vision; numerical analysis; inverse problems; deblurring; diffusion process; blurring; regularization; scale-space parameter ClassCodes B0290; B0290B; B6140C; C1250; C4100; C4110 Article Type Theoretical / Mathematical; Experimental Language English RecordType Conference ControlNo. 2977753 AbstractNos. B87062362; C87051420 ISBN or SBN 0 8186 0777 7 References 14 U.S. Copyright Clearance Center Code CH2465-3/87/0000-0527$01.00 Country Pub. USA date 1137 ------------------------------------------------------------ Author Clark, J.J.; Div. of Appl. Sci., Harvard Univ., Cambridge, MA, USA Title Singularities of contrast functions in scale space Source Proceedings of the First International Conference on Computer Vision (Cat. No.87CH2465-3); Part: London, UK; Part: 8-11 June 1987; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: C08; Washington, DC, USA; IEEE Comput. Soc. Press; xii+734; 1987; pp. 491-5 Abstract The process of detecting edges in a one-dimensional signal by finding the zeros of the second derivative of the signal can be interpreted as the process of detecting the critical points of a general class of contrast functions which are applied to the signal. It is shown that the concavity of the contrast function at a critical point is related to the classification of the associated edge as being phantom or authentic. The contrast of authentic edges is shown to decrease with filter scale, while the contrast of phantom edges is shown to increase with scale. It is shown that as the filter scale increases, an authentic edge must either turn into a phantom edge or join with a phantom edge and vanish. The points in the scale space at which these events occur are seen to be the singular points of the contrast function. Using ideas from singularity, or catastrophe, theory, one can show that the form of the scale map contours near these singular points is restricted to one of two basic types. The analysis of the behavior of the contrast function near a singularity also provides a proof of the property that scale map contours cannot be created as the filter scale increases Thesaurus catastrophe theory; filtering and prediction theory; pattern recognition Other Terms edge detection; contrast function singularities; contrast function concavity; singularity theory; catastrophe theory; pattern recognition; scale space; one-dimensional signal; authentic edges; phantom edges; filter scale; scale map contours ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2977746 AbstractNos. B87062355; C87051413 ISBN or SBN 0 8186 0777 7 References 17 U.S. Copyright Clearance Center Code CH2465-3/87/0000-0491$01.00 Country Pub. USA date 1137 ------------------------------------------------------------ Author Blostein, D.; Ahuja, N.; Coordinated Sci. Lab., Illinois Univ., Urbana, IL, USA Title Representation and three-dimensional interpretation of image texture: An integrated approach Source Proceedings of the First International Conference on Computer Vision (Cat. No.87CH2465-3); Part: London, UK; Part: 8-11 June 1987; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: C01; Washington, DC, USA; IEEE Comput. Soc. Press; xii+734; 1987; pp. 444-9 Abstract A perspective view of a slanted textured surface shows systematic changes in the density, area and aspect-ratio of texture elements. These apparent changes in texture element properties can be analyzed to recover information about the physical layout of the scene. However, in practice it is difficult to identify texture elements, especially in images in which the texture elements are partially occluded or are themselves texture at a finer scale. To solve this problem, it is necessary to integrate the extraction of texture elements with the recognition of scene layout. A method for recovering the orientation of textured surfaces, while simultaneously identifying texture elements, is presented. A multiscale region detector, based on measurements in a Delta /sup 2/G (Laplacian-of-Gaussian) scale-space, is used to construct a set of candidate texture elements. True texture elements are selected from the set of candidate texture elements by finding the planar surface that best predicts the observed properties of the candidate texture elements. Results are shown for a variety of natural textures, including waves, flowers, rocks, clouds and dirt clods Thesaurus pattern recognition; picture processing; surface texture Other Terms image texture representation; partial occlusions; scene layout recognition; feature extraction; Laplacian-of-Gaussian scale space; computer vision; three-dimensional interpretation; perspective view; slanted textured surface; orientation; multiscale region detector; waves; flowers; rocks; clouds; dirt clods ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2977740 AbstractNos. B87062349; C87051407 ISBN or SBN 0 8186 0777 7 References 14 U.S. Copyright Clearance Center Code CH2465-3/87/0000-0444$01.00 Country Pub. USA date 1137 ------------------------------------------------------------ Author Aviad, Z.; Dept. of Comput. Sci., Carnegie-Mellon Univ., Pittsburgh, PA, USA Title A discrete scale-space representation Source Proceedings of the First International Conference on Computer Vision (Cat. No.87CH2465-3); Part: London, UK; Part: 8-11 June 1987; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: B21; Washington, DC, USA; IEEE Comput. Soc. Press; xii+734; 1987; pp. 417-21 Abstract A discrete alternative to scale-space filtering is presented. The method provides for fast solutions to problems of spatial containment, filtering and matching, without using arbitrary parameters and smoothing of the input. The discrete space-scale representation is a hierarchical perceptual organization that has applications in computer vision research. Examples from actual implementation are provided Thesaurus computer vision; filtering and prediction theory Other Terms discrete scale-space representation; scale-space filtering; spatial containment; matching; arbitrary parameters; smoothing; hierarchical perceptual organization; computer vision ClassCodes B6140; C1250 Article Type Practical; Theoretical / Mathematical Language English RecordType Conference ControlNo. 2977736 AbstractNos. B87062134; C87051402 ISBN or SBN 0 8186 0777 7 References 14 U.S. Copyright Clearance Center Code CH2465-3/87/0000-0417$01.00 Country Pub. USA date 1137 ------------------------------------------------------------ Author Lyon, R.F.; Schlumberger Palo Alto Res., CA, USA Title Speech recognition in scale space Source Proceedings: ICASSP 87. 1987 International Conference on Acoustics, Speech, and Signal Processing (Cat. No.87CH2396-0); Par t: Dallas, TX, USA; Part: 6-9 April 1987; Sponsored by: IEEE; Translated in: G20; New York, NY, USA; IEEE; 4 vol. 2425; 1987; pp. 1265-8 vol.3 Abstract Scale-space filtering, proposed by A.P. Witkin (1984) for describing natural structure in one-dimensional signals, has been extended for application to segmentation and description of vector-valued functions of time, such as speech spectrograms. Scale-space segmentations of cochleagrams (spectrograms based on a computational model of the peripheral auditory system) have been experimentally applied to word recognition. Recognition using fixed-scale segmentations with finite-state word models and a Viterbi search has led to speaker-independent digit recognition accuracies of greater than 97%, about the same as the tests with nonsegmented cochleagrams Thesaurus filtering and prediction theory; speech recognition Other Terms scale space filtering; speech recognition; segmentation; vector-valued functions of time; speech spectrograms; cochleagrams; peripheral auditory system; word recognition; fixed-scale segmentations; finite-state word models; Viterbi search; speaker-independent digit recognition ClassCodes B6130; C1250C Article Type Experimental Language English RecordType Conference ControlNo. 2976912 AbstractNos. B87061729; C87051511 References 6 U.S. Copyright Clearance Center Code CH2396-0/87/0000-1265$01.00 Country Pub. USA date 1135 ------------------------------------------------------------ Author Withgott, M.; Bagley, S.C.; Lyon, R.F.; Bush, M.A.; Zerox Palo Alto Res. Center, CA, USA Title Acoustic-phonetic segment classification and scale-space filtering Source Proceedings: ICASSP 87. 1987 International Conference on Acoustics, Speech, and Signal Processing (Cat. No.87CH2396-0); Par t: Dallas, TX, USA; Part: 6-9 April 1987; Sponsored by: IEEE; Translated in: E18; New York, NY, USA; IEEE; 4 vol. 2425; 1987; pp. 860-3 vol.2 Abstract Scale-space filtering represents one method for automatically extracting both coarse and fine-grained units from the speech signal. The authors examine the acoustic-phonetic structure of segments obtained by scale-space filtering of cochleagrams, and report on the correspondences between scale-space segments which are automatically derived and hand-marked acoustic-phonetic segments. The major advantage of this segmentation is the flexibility of the data structure Thesaurus acoustic signal processing; filtering and prediction theory; speech analysis and processing; speech recognition Other Terms speech analysis; speech processing; acoustic-phonetic segment classification; speech recognition; scale-space filtering; speech signal; cochleagrams; data structure ClassCodes A4370; B6130; B6140 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2976811 AbstractNos. A87110719; B87061704 References 6 U.S. Copyright Clearance Center Code CH2396-0/87/0000-0860$01.00 Country Pub. USA date 1135 ------------------------------------------------------------ Author Sato, M.; Wada, T.; Kawarada, H.; Res. Lab. of Precision Machinery & Electron., Tokyo Inst. of Technol., Yokohama, Japan Title A hierarchical representation of random waveforms by scale-space filtering Source Proceedings: ICASSP 87. 1987 International Conference on Acoustics, Speech, and Signal Processing (Cat. No.87CH2396-0); Par t: Dallas, TX, USA; Part: 6-9 April 1987; Sponsored by: IEEE; Translated in: B19; New York, NY, USA; IEEE; 4 vol. 2425; 1987; pp. 273-6 vol.1 Abstract The authors introduce an analytic line, named the structure line, on the surface of generalized waveform f(x, sigma ). The structure line describes the relation of the convex and concave region of a generalized waveform. The structure line is defined by some derivatives of the generalized waveform, and has the same topological structure as a trinary tree. The properties of the structure line are discussed and some examples are shown. It is confirmed that the structure line is effective in representing waveforms hierarchically Thesaurus filtering and prediction theory; waveform analysis Other Terms convex region; hierarchical representation; random waveforms; scale-space filtering; analytic line; structure line; concave region; derivatives; trinary tree ClassCodes B0220; B6140 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2976666 AbstractNos. B87061938 References 9 U.S. Copyright Clearance Center Code CH2396-0/87/0000-0273$01.00 Country Pub. USA date 1135 ------------------------------------------------------------ Author Kass, M.; Witkin, A.; Terzopoulos, D.; Schlumberger Palo Alto Res., CA, USA Title Snakes: active contour models Source Proceedings of the First International Conference on Computer Vision (Cat. No.87CH2465-3); Part: London, UK; Part: 8-11 June 1987; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: B03; Washington, DC, USA; IEEE Comput. Soc. Press; xii+734; 1987; pp. 259-68 Abstract A snake is an energy-minimizing spline guided by external constraint forces and influenced by image forces that pull it toward features such as lines and edges. Snakes are active contour models; they lock into nearby edges, localizing them accurately. Scale-space continuation can be used to enlarge the capture region surrounding a feature. Snakes provide a unified account of a number of visual problems, including detection of edges, lines, and subjective contours, motion tracking, and stereo matching. The authors have used snakes successfully for interactive interpretation, in which user-imposed constraint forces guide the snake near features of interest Thesaurus computer vision; pattern recognition Other Terms active contour models; snake; energy-minimizing spline; external constraint forces; image forces; lines; edges; capture region; detection of edges; motion tracking; stereo matching; interactive interpretation ClassCodes C1250 Article Type Practical; Experimental Language English RecordType Conference ControlNo. 2967219 AbstractNos. C87051386 ISBN or SBN 0 8186 0777 7 References 20 U.S. Copyright Clearance Center Code CH2465-3/87/0000-0259$01.00 Country Pub. USA date 1137 ------------------------------------------------------------ Author Johansen, P.; Skelboe, S.; Grue, K.; Andersen, J.D.; Dept. of Datalogy, Copenhagen Univ., Denmark Title Representing signals by their top-points in scale space Source Eighth International Conference on Pattern Recognition. Proceedings (Cat. No.86CH2342-4); Part: Paris, France; Part: 27- 31 Oct. 1986; Sponsored by: IEEE; Int. Assoc. Pattern Recognition; Translated in: B25; Washington, DC, USA; IEEE Comput. Soc. Press; xxxvi+1300; 1986; pp. 215-17 Abstract In 1983, A. Witkin introduced the fingerprint of a function as a set F of points in scale space, where scale space is the plane. Fingerprints are calculated by convolving the function with a Gaussian filter with continuously varying standard deviation. Within defined the top-points of the signal as points in scale space where F has a horizontal tangent. It is proved that periodic, bandlimited functions are defined up to a multiplicative constant by their top-points, if this concept is properly generalized. The uniqueness theorem may be regarded as a sampling theorem for signals in the scale space Thesaurus filtering and prediction theory; signal processing Other Terms signal representation; signal processing; scale space; top- points; multiplicative constant; sampling theorem ClassCodes B6140; C1260 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2933821 AbstractNos. B87047850; C87037981 ISBN or SBN 0 8186 0742 4 References 16 U.S. Copyright Clearance Center Code CH2342-4/86/0000-0215$01.00 Country Pub. USA date 1128 ------------------------------------------------------------ Author Witkin, A.; Terzopoulos, D.; Kass, M.; Schlumberger Palo Alto Res., CA, USA Title Signal matching through scale space Source Proceedings AAAI-86: Fifth National Conference on Artificial Intelligence; Part: Philadelphia, PA, USA; Part: 11-15 Aug. 1986 ; Sponsored by: American Assoc. Artificial Intelligence; Univ. Pennsylvania; Translated in: F13; Menlo Park, CA, USA; American Assoc. Artificial Intelligence; 2 vol. xxxii+1165; 1986; pp. 714-19 vol.1; Available from: Morgan Kaufmann Publishers Inc., Los Altos, CA, USA Abstract Given a collection of similar signals that have been deformed with respect to each other, the general signal matching problem is to recover the deformation. The authors formulate the problem as the minimization of an energy measure that combines a smoothness term and a similarity term. The minimization reduces to a dynamic system governed by a set of coupled, first-order differential equations. The dynamic system finds an optimal solution at a coarse scale and then tracks it continuously to a fine scale. Among the major themes in recent work on visual signal matching have been the notions of matching as constrained optimization, of variational surface reconstruction, and of coarse-to-fine matching. The solution captures these in a precise, succinct, and unified form. Results are presented for one- dimensional signals, a motion sequence, and a stereo pair Thesaurus picture processing; signal processing Other Terms image processing; scale space; signal matching; visual ClassCodes C1250 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2920186 AbstractNos. C87037546 ISBN or SBN 0 934613 13 3 References 17 Country Pub. USA date 1126 ------------------------------------------------------------ Author Carlotto, M.J.; Anal. Sci. Corp., Reading, MA, USA Title Histogram analysis using a scale-space approach Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.PAMI-9, no.1; Translated in: A10; Jan. 1987; pp. 121-9 Abstract A new application of scale-space filtering to the classical problem of estimating the parameters of a normal mixture distribution is described. The technique involves generating a multiscale description of a histogram by convolving it with a series of Gaussians of gradually increasing width (standard deviation), and marking the location and direction of the sign change of zero-crossings in the second derivative. The resulting description, or fingerprint, is interpreted by relating pairs of zero-crossings to modes in the histogram where each mode or component is modeled by a normal distribution. Zero-crossings provide information from which estimates of the mixture parameters are computed. These initial estimates are subsequently refined, using a iterative maximum likelihood estimation technique. Varying the scale or resolution of the analysis allows the number of components used in approximating the histograms to be controlled Thesaurus filtering and prediction theory; parameter estimation; picture processing Other Terms histogram analysis; parameter estimation; picture processing; scale-space filtering; normal mixture distribution; zero- crossings; fingerprint ClassCodes B6140C; C1220; C1250 Article Type Practical; Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 2901022 AbstractNos. B87039898; C87031838 ISSN 01628828 References 14 U.S. Copyright Clearance Center Code 0162-8828/87/0100-0121$01.00 Country Pub. USA date 1132 ------------------------------------------------------------ Author Blake, A.; Zisserman, A.; Papoulias, A.V.; Dept. of Comput. Sci., Edinburgh, UK Title Weak continuity constraints generate uniform scale-space descriptions of plane curves Source ECAI '86. 7th European Conference on Artificial Intelligence. Proceedings; Part: Brighton, UK; Part: 21-25 July 1986; Translated in: C01; London, UK; Conference Services; 2 vol. (597+xxxii+187); 1986; pp. 518-28 vol.1 Abstract Scale-space filtering is a recently developed technique, both powerful and general, for segmentation and analysis of signals. Asada and Brady (1984) have amply demonstrated the value of scale- space for description of curved contours from digitised images. Weak continuity constraints furnish novel, powerful, nonlinear filters, to use in place of gaussians, for scale-space filtering. This has some striking advantages. First, scale-space is uniform, so that tracking across scale is a trivial task. Structure need not be preserved to indefinitely fine scale; this leads to an enrichment of the concept of scale-a rounded corner, for example, can be represented as a discontinuity at coarse scale but smooth at fine scale. And, finally, boundary conditions at ends of curves are handled satisfactorily-it is as easy to analyse open curves as closed ones Thesaurus boundary-value problems; computational geometry; computerised pattern recognition; curve fitting; dynamic programming; filtering and prediction theory Other Terms weak continuity constraints; signal analysis; signal segmentation; nonpreserved structure; plane curves; digitised images; nonlinear filters; scale-space filtering; boundary conditions ClassCodes C1180; C1250; C1260; C4130 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2832019 AbstractNos. C87014666 References 17 Country Pub. UK date 1125 ------------------------------------------------------------ Author Goshtasby, A.; Mokhtarian, F.; Mackworth, A.; Dept. of Comput. Sci., Kentucky Univ., Lexington, KY, USA Title Comments on 'Scale-based description and recognition of planar curves and two-dimensional shapes' (with reply) Source IEEE Transactions on Pattern Analysis and Machine Intelligence; IEEE Trans. Pattern Anal. Mach. Intell. (USA); vol.PAMI-8, no.5; Translated in: A09; Sept. 1986; pp. 674-5 Abstract The commenter argues that in the above paper (ibid., vol.PAMI-8, p., 34-43, Jan. 1986), which describes a technique for registration of a Landsat image with a map, if the scale of details in the map and the Landsat image are different, scale- space images of contours in the map and scale-space images of region boundaries in the image will not match reliably. The authors of the paper rebut the commenter's argument Thesaurus computerised pattern recognition; computerised picture processing ; remote sensing Other Terms scale based description; pattern recognition; scale based recognition; 2D shapes; picture processing; image processing; pattern recognition; planar curves; Landsat image; scale-space images ClassCodes B6140C; C1250 Article Type Theoretical / Mathematical Coden ITPIDJ Language English RecordType Journal ControlNo. 2800963 AbstractNos. B87007830; C87005113 ISSN 01628828 References 2 U.S. Copyright Clearance Center Code 0162-8828/86/0900-0674$01.00 Country Pub. USA date 1127 ------------------------------------------------------------ Author Hummel, R.A.; Courant Inst. of Math. Sci., New York Univ., NY, USA Title Representations based on zero-crossings in scale-space Source Proceedings CVPR '86: IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.86CH2290-5); Part : Miami Beach, FL, USA; Part: 22-26 June 1986; Sponsored by: IEEE; Translated in: B05; Washington, DC, USA; IEEE Comput. Soc. Press; viii+676; 1986; pp. 204-9 Abstract Using the heat equation to formulate the notion of scale-space filtering, the author shows that the evolution property of level- crossings in scale-space is equivalent to the maximum principle. He briefly discusses filtering over bounded domains. He then considers the completeness of the representation of data by zero- crossings, and observes that for polynomial data, the issue is solved by standard results in algebraic geometry. For more general data, it is argued that gradient information along the zero-crossings is needed, and that although such information more than suffices, the representation is still not stable. The author gives a simple linear procedure for reconstruction of data from zero-crossings and gradient data along zero-crossings in both continuous and discrete scale-space domains Thesaurus filtering and prediction theory; maximum principle; polynomials; signal processing Other Terms zero-crossings; heat equation; scale-space filtering; evolution property; maximum principle; bounded domains; completeness; polynomial data; algebraic geometry; linear procedure; gradient data; discrete scale-space domains ClassCodes B0290F; B6140; C1260; C4130 Article Type Theoretical / Mathematical Language English RecordType Conference ControlNo. 2790139 AbstractNos. B87002549; C87000689 ISBN or SBN 0 8186 0721 1 References 20 U.S. Copyright Clearance Center Code CH2290-5/86/0000-0204$01.00 Country Pub. USA date 1124 ------------------------------------------------------------ Author Richards, W.; Dawson, B.; Whittington, D.; Natural Computation Group, MIT, Cambridge, MA, USA Title Encoding contour shape by curvature extrema Source Journal of the Optical Society of America A (Optics and Image Science); J. Opt. Soc. Am. A, Opt. Image Sci. (USA); vol.3, no.9; Translated in: B03; Sept. 1986; pp. 1483-91 Abstract Curvature extrema provide significant information about the shape of an image contour, such as a silhouette, and are the basis for the Hoffman-Richards codon representation for shape. This representation based on curvature easily translates into a binary string that will describe the abstract shape of any smooth image curve. The computation of the basic shape primitives requires dealing with two even-pervasive problems: contour noise and scale. The authors show how contour noise can be estimated given knowledge of the shape of the filter used to compute curvature from the edge list of the contour. To handle the scale problem, the authors use an adaptation of Witkin's scale space. The authors algorithm differs from Witkin's by using a notion of parts to set criteria for significant structures Thesaurus computer vision; encoding Other Terms computer vision; machine vision; shape representation; encoding ; curvature extrema; image contour; silhouette; Hoffman- Richards codon representation; basic shape primitives; contour noise ClassCodes B6120B; B6140C; C1250 Article Type Theoretical / Mathematical Coden JOAOD6 Language English RecordType Journal ControlNo. 2784401 AbstractNos. B87002576; C87000503 ISSN 07403232 References 32 U.S. Copyright Clearance Center Code 0740-3232/86/091483-09$02.00 Country Pub. USA date 1127