Date: 11 Jul 90 09:51:03-PST
From: Vision-List moderator Phil Kahn <Vision-List-Request@ADS.COM>
Errors-to: Vision-List-Request@ADS.COM
Reply-to: Vision-List@ADS.COM
Subject: Vision-List delayed redistribution
To: Vision-List@ADS.COM

Vision-List Digest	Wed Jul 11 09:51:04 PDT 90

 - Send submissions to Vision-List@ADS.COM
 - Send requests for list membership to Vision-List-Request@ADS.COM

Today's Topics:

 Help ! Rotated molecules & rotating physics student
 canonical vision book list
 Summary of Responses Received to generate Needle Diagrams
 Camera Mount
 Ideas Needed for Manned Exploration of Moon and Mars
 Character Recognition Bibliography?
 OCR Refs/Software?
 Job vacancies

----------------------------------------------------------------------

Date: 10 Jul 90 10:34:49 GMT
From: uh311ae@LRZSUN4_7.lrz.de (Henrik Klagges)
Subject: Help ! Rotated molecules & rotating physics student
Keywords: Correllation filters, rotated pattern recognition, GA
Organization: LRZ, Bavarian Academy of Sciences, W. Germany

Hi, netwise,
I need help !

The problem:

A Scanneling Tunnel Microscope (STM) produces a picture of a flat surface 
covered with macromolecules, forming a loose grid or just being scattered
around. For simplicity, it is assumed that there exist only 3 free para-
meters, namely two translatoric and one rotational for the molecules. 
A single molecule gives a noisy image, so it is desired to combine many 
individual molecule-images into a single one. To accomplish this task
several ways might be possible:

1) Have a clever program walking over the image and saying 'Wow ! That's a
   molecule turned foo degrees and translated bar units, let's add it to our
   data base !' (Ugh).

2) Run a fantastic correllation filter (fcf:) over the image that is able 
   to recognize the correllation between any (!) rotated and x,y-displaced 
   structures and amplify those structures (Does this exist yet ? Does one
   exist that matches all affine transformations ?).

3) If that is too much, select a "good" molecule, calculate its turned image
   for each degree, move over the whole image and try to match these 360 
   turns with the image and mark this place as occupied (Calculate 'n crunch 
   for ever ?).

4) Make a FF- or Hartley- or another integral transform of the image. That means
   no spatial parameters anymore, and then turn and match the transformed image 
   on itself, correllate, amplify, re-transform (Who knows if that works !).

Questions:

1) Who knows about a fcf ?

2) Which methods are suited for the task of matching two images that are 
   rotated and/or linearly displaced against each other ? (I heard about
   a genetic algorithm from Fitzgerald, Grevenstette et al.).

3) How do You walk over a surface and recognize noisy molecules ?

4) ANY hint or comment desperately welcome !


Thanks a lot

Henrik Klagges

STM group at LMU Munich
EMail: 	uh311ae@LRZSUN4_7.lrz.de
SMail:  Stettener Str. 50, 8210 Prien, FRG


------------------------------

Date: Tue, 10 Jul 90 12:39:23 +0200
From: ronse@prlb.philips.be
Subject: canonical vision book list

I received many references of books on vision, but of only one on image
analysis (Serra's first volume). So here is my "canonical" list of
fundamental vision books. Please send me also the references for your
preferred books on image processing and analysis.

D.H. Ballard & C.M. Brown: "Computer Vision", Prentice-Hall, 1982.

B.K.P. Horn: "Robot Vision", MIT Press, Cambridge, Mass., USA, 1986.

T. Kanade : "Three Dimensional Machine Vision", Kluwer Ac. Press.

J.J. Koenderink: "Solid Shape", MIT Press, Cambridge, Mass., USA., 1990.

M.D. Levine: "Vision in Man and Machine", McGraw-Hill, New York, USA,
1985.

D. Marr: "Vision", W.H. Freeman & Co., San Francisco, CA, 1982.

P.H. Winston (ed.): "The Psychology of Computer Vision", McGraw-Hill,
1975.

The books by Ballard & Brown and by Horn are "winners". Several people
suggested them.

Christian Ronse

Internet:			ronse@prlb.philips.be
BITNET:				ronse%prlb.philips.be@cernvax

------------------------------

Date: Fri, 6 Jul 90 17:09:41 EDT
From: ramamoor@cmx.npac.syr.edu (Ganesh Ramamoorthy)
Subject: Summary of Responses Received to generate Needle Diagrams

>From: jonh@tele.unit.no
Subject: Needle Diagrams

If you use matlab, ver. 3.5 you can use the command "quiver"
to generate nice needle diagrams. Unfortunately, ver 3.5 of matlab
may still not be available on SUN. However it is available for
VAX. I have taken the m-file "quiver.m" from Matlab ver. 3.5
running on VAX and used it with old matlab versions on SUN. 
This works fine. If you would like to try this out and have
problems getting hold of "quiver.m", let me know and I will
mail you a copy.

>From: Keith Langley <kl@robots.oxford.ac.uk>
Subject: needles

I did it the easy way by looking at FIG format for lines
and piping out a needle diagram that way.

>From: hmueller@wfsc4.tamu.edu (Hal Mueller)
Subject: Needle Diagrams

The graphics package called DISSPLA, sold by Computer Associates (formerly
sold by ISSCO), has an extra cost option for an automatic code generator.
One of the things this code generator produces is needle plots.  DISSPLA
runs on a Sun 3, and I would presume also on a Sun 4.

Reach them at (800) 645-3003.  The package is expensive, fairly difficult
to learn, is extremely powerful, and is as close to bug free as any
commercial product I've ever seen.

>From: oskard@vidi.cs.umass.edu
Subject: re: Programs for generating Needle Diagrams

	Hi.  I don't know if you've found any programs for displaying
flow fields yet, but we use KBVision here and it has a system called
the image examiner that displays flow fields among other things.
Anyway, their address is:

	Amerinex Artificial Intelligence Inc.
	274 N. Pleasant St.
	Amherst, MA 01002

	413-256-8941

>From: johnston@george.lbl.gov (Bill Johnston)
Subject: flow field diagramw

The NCAR (National Center for Atmospheric Research, Boulder CO)
graphics library has several routines for displaying flow
fields. It sits on top of GKS, so any GKS package that supports
a Sun should allow you to use the NCAR package.

------------------------------

Date: Sun,  8 Jul 90 12:38:29 -0400 (EDT)
From: "Ethan Z. Evans" <ee0i+@andrew.cmu.edu>
Subject: Camera Mount

Vision Wizards:
	I need to find a 2 degree of freedom camera mount (rotation and pitch).
 The mobile platform I work on is only about a foot tall, and the top is
cluttered with various hardware etc.  Thus to give the robot a view of
its world, we need to set the camera up on a pole.  Once its up their it
needs to be able to turn around and either look down at what the arm is
doing, or out at where the base is going.  Point is, I don't have the
slightest idea where to look for such a device.  If anyone could give me
a starting point as to where to get such a mount, especially one easily
controlled through a PC parallel or serial port, I would be most
greatful.
	
Thanks in advance,
	Ethan Evans
	ee0i@andrew.cmu.edu
	
Disclaimer:  I'm the lab rat, how could *I* have an opinion?

[ Please post responses to the List
		phil...	]

------------------------------

Date: 9 Jul 90 07:15:53 GMT
From: guyton%randvax.UUCP@usc.edu (Jim Guyton)
Subject: Ideas Needed for Manned Exploration of Moon and Mars
Keywords: PROJECT OUTREACH
Organization: Rand Corp., Santa Monica, Ca.

        PROJECT OUTREACH

        Ideas Needed for Manned Exploration of Moon and Mars.

        NASA is seeking innovative approaches to mission concepts and
        architectures, as well as technologies that could cut costs and
        improve mission schedule and performance.

        The RAND Corporation will provide an independent assessment
        of all suggestions.

        The procedure for submitting ideas is simple. For an information
        kit call 1-800-677-7796.  Call now.  The deadline for submissions
        is August 15, 1990.

------------------------------

Date: Mon, 9 Jul 90 19:56:22 EDT
From: atul k chhabra <Atul.Chhabra@UC.EDU>
Subject: Character Recognition Bibliography?

I am looking for a character recognition bibliography. I am interested in all
aspects of character recognition, i.e., preprocessing and segmentation,
OCR, typewritten character recognition, handwritten character recognition,
neural network based recognition, statistical and syntactic recognition,
hardware implementations, and commercial character recognition systems.

If someone out there has such a bibliography, or something that fits a part
of the above description, I would appreciate receiving a copy. Even if you
know of only a few references, please email me the references.

Please email the references or bibliography to me. I will summarize on the 
vision-list. Thanks,

Atul Chhabra
Department of Electrical & Computer Engineering
University of Cincinnati, ML 030
Cincinnati, OH 45221-0030

Phone: (513)556-6297
Email: achhabra@uceng.uc.edu

------------------------------

Date: Wed, 11 Jul 90 14:57:00 +0200
From: nagler@olsen.ads.com (Robert Nagler)
Subject: OCR Refs/Software?
Organization: Olsen & Associates, Zurich, Switzerland
Keywords: OCR, Pattern Recognition
Status: R

Could someone send me a good reference(s) on OCR/Pattern Recognition?
Pointers to PD software (source) would be nice, too.   Thanks.

Rob nagler@olsen.uu.ch

------------------------------

Date: Wed, 4 Jul 90 11:29:00 WET DST
Subject: job vacancy
From: J.Illingworth@ee.surrey.ac.uk

    UNIVERSITY OF SURREY: Dept of Electronics and Electrical Engineering.

    *********************************************************************

        RESEARCH FELLOW IN COMPUTER VISION AND IMAGE PROCESSING

    *********************************************************************



  Research Fellows are required for projects in Computer Vision. 
  The projects are concerned with the following problems: 

    *  robust 2D and 3D shape representation and  analysis; 
    *  high-level scene interpretation; 
    *  automatic inspection of loaded printed circuit boards; 
    *  relaxation labelling and neural net computation in vision 
       by associative reasoning. 

  The projects will be carried out within an active Vision, Speech and 
  Signal Processing research group which comprises about 35 members. 
  The group has extensive computing resources including SUN Sparc stations, 
  VAX and Masscomp computers as well as specialised image processing 
  facilities.

  The successful candidates will be required to develop, implement in 
  software  and experimentally evaluate computer vision algorithms. 
  Applicants for these posts should have a degree in mathematics, 
  statistics, electronics, computer science, artificial intelligence or 
  physics. Previous experience in computer vision, image analysis, 
  knowledge based methods or pattern recognition will be an advantage. 
  One of the posts may be reserved for an applicant able to provide 
  hardware and software support for the Group across projects. 

  The appointments will be initially for two years with a salary in the range 
  10458 -16665 pa (under review) depending upon age, qualifications and 
  experience, with superannuation under USS conditions. Applications in the 
  form of a curriculum vitae ( 3 copies) including the names and addresses 
  of two referees should be sent to the Personnel Office (JLG), University 
  of Surrey, Guildford, Surrey GU2 5XH by  30 June  1990 quoting reference 


  Further information may be obtained from Dr J Kittler, Department of 
  Electronic and Electrical Engineering on (0483) 509294 or from 
  Dr J Illingworth on (0483) 571281 ext. 2299. 



    PROJECT DESCRIPTION

 DETECTION OF LINEAR FEATURE IN IMAGE DATA

 The aim of the project is to investigate and develop the Hough 
 transform and associated pre-processing and postprocessing techniques with 
 the application to the problem of detecting  linear image features in noisy 
 and cluttered background of changing polarity. The emphasis of the approach
 will be on statistical hypothesis testing and robust estimation methods.
 An important component of the research will be to develop the theory and 
 methodology for the design of post processing filters to enhance the Hough 
 transform performance. The problem of detecting higher level features such as 
 corners and parallels using the same framework will also be considered.


 LOCATION AND INSPECTION WITH RANGE DATA

 The project is concerned with the problem of segmenting depth image data into 
 parametric surfaces. Robust hypothesis testing methods of which the Hough 
 transform is just one example will be investigated in this context.
 The research issues to be addressed include the problems of surface 
 parameterisation, efficient transform calculation and reliable transform 
 space analysis. Other approaches to range data segmentation 
 such as energy minimisation methods and knowledge based methods will be 
 investigated. 


 AUTOMATIC INSPECTION OF LOADED PRINTED CIRCUIT BOARDS 

 The aim of the project is to investigate and develop machine vision 
 techniques for advanced automated inspection of loaded printed circuit 
 boards (PCBs) and of surface mounted ceramic substrates with the view to
 enable real time fault correction to increase production yields,
 maintain and optimise product quality and to maximize manufacturing process 
 control capabilities. A particular emphasis will be on two aspects of the 
 inspection task: i) the use of 3D sensor data to address the problem of 
 inspecting soldered joints, active device pins, component leads and 
 the mounting of special devices, ii) development of knowledge based 
 approaches to guide and control the above surface  based inspection problem
 and to verify component identity. The strategic objectives of the project 
 include the following: Development of 3D scene modelling and surface 
 segmentation methods specifically in the context of the loaded PCB inspection 
 domain; Development of surface inspection approaches; Representation and 
 application of geometric, attribute and relational models of objects and 
 their arrangements in the application domain of electronic system assembly.  
 The research problems to be addressed in order to meet these objectives are 
 generic in nature. It is anticipated therefore that the research results 
 will have bearing on other application areas of computer vision.

 The proposed research will advance the state of the art in automatic loaded 
 PCB inspection by:
 1 evaluating  existing 3D surface segmentation methods in the 
   context of loaded PCB inspection,
 2 developing robust surface modelling methods,
 3 providing techniques and algorithms for surface inspection, 
 4 enhancing the methods of component identification and 
   verification, and
 5 developing inspection strategies that will allow full integration 
   of automatic printed circuit assembly, inspection and rework.

 VISION BY ASSOCIATIVE REASONING

 The project is concerned with the study of relaxation labelling processes
 in the computer vision context. The aim of the research will be to develop
 and apply existing probabilistic and discrete relaxation algorithms
 to image interpretation problems at intermediate levels of processing
 where the prior world knowledge may most naturally be specified in terms of 
 explicit rules. It will be necessary to develop a suitable interface that 
 facilitates the conversion of such rules into a form that can be used 
 directly by the evidence combining scheme employed in the relaxation 
 process. The work will also involve the development of evidence combining 
 methods for multilevel relaxation, development of hierarchical models
 and corresponding hierarchical relaxation processes. The relationship
 of relaxation processes and neural net computation will be investigated.

------------------------------

End of VISION-LIST
********************
