Date: 15 Mar 90 09:39:46-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	Thu Mar 15 09:39:46 PDT 90

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

Today's Topics:

 Vector correlation
 Use of Artificial Neural Networks in Image Processing
 Additions to List of IP Source Code Packages
 Australian AI Conference
 Call for discussion:  comp.robotics

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

Date: Tue, 13 Mar 90 04:00:21 GMT
From: us214777@mmm.serc.3m.com (John C. Schultz)
Subject: vector correlation
Organization: 3M - St. Paul, MN  55144-1000 US

A friend gave me a copy of an article from the Jan 1990 (p 138-9) Photonics
Spectra on "Vector Correlation" which was the first I have heard of the
concept of vector correlation.

As I understand what the author was talking about, you use a high-pass filter
such as a Sobel to determine edge magnitude and gradient (the article only
discussed 4 angles vs. Sobel's 8).  The correlation for object location can
then be done much more robustly wrt lighting variations by correlating the 4
(or 8 for Sobel?)  images for each direction vector and summing the resulting
correlation images.

The advantages of this approach would seem to be:
1. insensivity to light level even as compared to normalized correlation
2. greater location precision since the object location is completely
   determined by its edge location(s)

The disadvantage is the computational complexity - what was one correlation
has suddenly become 4 (or 8 in the case of Sobel?).

My questions:

Why vector correlation?  Seems to me this is just a fancy name for object edge
correlation. 

Does anyone know have any references for this technique?  Possibly under a
different name?  The author was from Applied Intelligent Systems Inc. and
neglected to include any references :-).

Anyone have any experience with this technique?  Any code they would be
willing to share?

As a final note, I think this is what Intelledex uses internally to their
"turn-key" vision system and they do get 1/10 pixel location precision under a
lot of variable and/or poor lighting conditions.

John C. Schultz                   EMAIL: jcschultz@mmm.3m.com
3M Company,  Building 518-01-1     WRK: +1 (612) 733 4047
1865 Woodlane Drive, Dock 4
Woodbury, MN  55125

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

Date: Wednesday, 14 Mar 1990 13:37:03 EST
From: m20163@mwvm.mitre.org (Nahum Gershon)
Subject: Use of Artificial Neural Networks in Image Processing

I am looking for references on the use of Artificial Neural Networks
in image processing and also in biomedical imaging.  Does anyone have
any information?

  * * Nahum

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

Date: Tue, 13 Mar 90 12:09:16 PST
From: Scott E. Johnston <johnston@odin.ads.com>
Subject: Additions to List of IP Source Code Packages

In a recent posting to the vision-list I listed packages of image
processing source code, available in the public domain or for a
one-time license.  I inadvertently left out the HIPS software package
developed by Michael Landy.  My apologies to Professor Landy.  Here is
the information on HIPS.  In addition I received information about
a package called XVision available from the University of New
Mexico.  Once again, I welcome any additions or corrections to this list. 

Scott E. Johnston
johnston@ads.com
Advanced Decision Systems, Inc.
Mountain View, CA  94043

========================================================================

HIPS	

Contact:	SharpImage Software  
		P.O. Box 373		
		Prince St. Station
		NY, NY  10012

		Michael Landy (212) 998-7857 
		landy@nyu.nyu.edu

Description:

HIPS consists of general UNIX pipes that implement image processing
operators.  They can be chained together to implement more complex
operators.  Each image stores history of transformations applied.
HIPS is available, along with source code, for a $3000 one-time
license fee.

HIPS supports the following:
	- simple image transformations
	- filtering
	- convolution
	- Fourier and other transforms
	- edge detection and line drawing manipulation
	- image compression and transmission
	- noise generation
	- image pyramids 
	- image statistics
	- library of convolution masks
	- 150 programs in all

========================================================================

XVision

Contact:	John Rasure
		Dept. of EECE
		University of New Mexico
		Albuquerque NM 87131
		505-277-1351
		rasure@bullwinkle.unm.edu


XVision is a C-based system developed at the University of New Mexico.

It includes:	1) an image processing library of 150 algorithms from early
		processing to classification and shape analysis
		2) a graphical form/menu interface and a command line interface
		3) a visual language for configuring complex image processing
		pipelines
		4) an extensive 2d and 3d plotting capability
		5) an interactive image analysis capability
		6) code generators for generating the command line user 
		interface and the X windows user interface C code.

The system runs on most UNIX systems and they have a limited number of
licenses that they can give at no cost.


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

Date: Wed, 14 Mar 90 13:24:11 +0800
From: les@wacsvax.cs.uwa.oz.au (Les Kitchen)
Subject: Australian AI Conference

                                CALL  FOR  PAPERS

          4th Australian  Joint  Conference  on Artificial  Intelligence
              AI'90
          21-23 November, 1990
          Hyatt Regency, Perth, Western Australia

                       Everyday AI - New Tools for Society
                 This conference  is a  major regional  forum for
                 the  presentation  of  recent  research  on  the
                 theory and practical  applications of Artificial
                 Intelligence.    It   acts  as  a   catalyst  to
                 stimulate further  research  and cooperation  in
                 this important area  within the  Australasia and
                 Indian-Pacific region. The theme  of this year's
                 conference aims  to  encourage  and  promote  AI
                 techniques  and   tools  for   solving  everyday
                 problems.

          Topics Include (but not limited to):

          * Logic and Reasoning
          * Knowledge Representation and Acquisition
          * Machine Learning
          * Artificial Neural Networks  
  ====>   * Computer Vision and Robotics
          * Natural Language and Speech Recognition
          * Expert Systems and development tools
          * Applied AI in Civil,  Electrical, Electronic, and Mechanical
              Engineering
          * Knowledge Engineering in Business Applications
          * Applications in Government and Mining

          Criteria for acceptance
          This conference  welcomes  high quality  papers  which have  a
          significant contribution to  the theory   or practice  of A.I.
          Papers in  the  application  areas  will  be  judged by  their
          novelty in  the application,  its formulation,  application of
          new A.I.  techniques,  and  the  success  of  the  application
          project.

          Requirement for submission
          Authors must submit four copies of  their full papers to AI'90
          Programme Committee by 11th May 1990.    Submissions after the
          deadline may be returned without being opened. Notification of
          acceptance and format of the camera  ready copy will be posted
          by the 27th  July 1990. The  camera ready final  paper will be
          due on 24th August 1990.

          Paper Format for Review
          The paper should be about  5000 words in length.  It should be
          at least one and  a half spacing and  clearly legible. Authors
          should try to limit their paper to not  more than 15 pages not
          including diagrams.  Each  paper  must  include  a  title,  an
          abstract about 100 words, but no other identifying marks. The
          abstract of  100  words with  the  title,  authors names,  and
          correspondence address  should accompany  the submission  on a
          separate page.

          Publication
          All papers accepted in the conference will be published in the
          conference  proceedings.  Following  the   tradition  of  this
          conference, effort  will  also  be  made  to publish  selected
          papers from the conference in book form for wider circulation.

          Submission Check List
          When submitting your paper, please include the following: Name
          of contact, postal  address, telephone  (with country  code if
          applicable), fax number,  e-mail address,  FOUR copies  of the
          paper, an abstract, and a biographical note of the authors.

          Submit papers to:
          AI'90 Programme Committee
          c/o Department of Computer Science,
          University of Western Australia,
          Nedlands, W.A. 6009,
          AUSTRALIA


          Enquiries to:
                Dr. C.P.Tsang, AI'90 Programme Chair,
                Tel: +61-9-380-2763
                Fax: +61-9-382-1688
                email: ai90paper@wacsvax.oz.au

          This  conference  is  sponsored  by  the  Australian  Computer
          Society  through  the  National  Artificial  Intelligence  and
          Expert Systems Committee.

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

Date: 9 Mar 90 00:18:30 GMT
From: ttidca.TTI.COM!hollombe%sdcsvax@ucsd.edu (The Polymath)
Subject: Call for discussion:  comp.robotics
Organization: Citicorp/TTI, Santa Monica

[ I post this for your information only.  Please direct responses to 
  the poster of this message. 
		phil...	]


The broad distribution of this proposal should give some idea as to why I
consider the creation of this group appropriate.  The subject of robotics
can draw on information from all of these groups and more, yet none is
particularly appropriate to it.  There is no one group I could go to with
a specific robotics problem with a high probability of finding anyone who
even has the same problem, let alone a solution.  Rather than broadcast
such questions to the net, I'd like to see a dedicated group formed.

I therefore propose a new group:

     Name:     comp.robotics

     Moderation:  Unmoderated

     Charter:  The discussion and exchange of information on the practical
	       aspects of real-world robots and their applications --
               industrial, personal and experimental.

I put in the "real-world" qualification deliberately to exclude
discussions of positronic brains, R2D2 and who, if anyone, was really
inside Robbie in "Forbidden Planet."  I suggest that Asimov's laws are also
best left to a more philosophically or socially oriented group.

For those interested in leading edge research, AI, machine vision, etc. a
sci.robotics group might be more appropriate and can also be discussed at
this time.  I don't think the two groups are mutually exclusive. (i.e.:
Creation of one doesn't necessarily remove the need for the other).

Follow-ups are directed to news.groups

The Polymath (aka: Jerry Hollombe, hollombe@ttidca.tti.com)  Illegitimis non
Citicorp(+)TTI                                                 Carborundum
3100 Ocean Park Blvd.   (213) 450-9111, x2483
Santa Monica, CA  90405 {csun | philabs | psivax}!ttidca!hollombe


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

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