From vision@deimos.ads.com Wed Jul  6 02:09:31 1988
Received: from ads.com by deimos.ads.com (5.59/1.11)
	id AA00353; Wed, 6 Jul 88 02:09:25 PDT
Received: from deimos.ads.com by ads.com (5.59/1.14)
	id AA13266; Wed, 6 Jul 88 02:13:11 PDT
Received: by deimos.ads.com (5.59/1.11)
	id AA01826; Sat, 11 Jun 88 10:09:28 PDT
Message-Id: <8806111709.AA01826@deimos.ads.com>
Date: 11 Jun 1988 09:41:58-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
Status: RO

Vision-List Digest	Sat Jun 11 09:41:58 PDT 1988

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

Today's Topics:

 -- Vision Guided Robotics
 -- ICIP'89 : 1989 IEEE International Conference on Image Processing
 -- Reply to request on DataCube
 -- Quota Studentships in Computer Science
 -- Detection of pipe deterioration
 -- AI conference
 -- Source code for STEREOPSIS algorithms
 -- Multiple copies of Vision List Digest

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

Date: Wed, 25 May 88 18:21:00 PDT
From: pkahn@meridian.ads.com (Phil Kahn)
Subject: vision guided robotics


[This was posted on comp.ai,sci.electronics,rec.ham-radio,rec.video,
 but I thought it might be of general interest.
	- pk ]


	Date: 12 May 88 04:04:02 GMT
	From: jbn@glacier.STANFORD.EDU (John B. Nagle)
	Subject: TV systems for mobile robots
	Keywords: TV vision robot RF transmitter
	Organization: Stanford University

	
	      I'd like to hear about experience with various cameras
	and radio links used with mobile robots.  I'm interested in
	units suitable for a small, high-speed vehicle in which the
	vision processing is offboard. 
	
	      The ideal device, as pointed out by Russell Anderson in "A Robot
	Ping-Pong Player", is a CCD frame-transfer image sensor, since
	with such devices the entire frame is acquired as a unit and
	no artifacts of the scanning process appear in the image.  Examples of
	such parts are the Sanyo LC99xx series.  (The Fisher-Price Toy
	Camcorder and Lionel  Loco-Vision use the LC9943, a low-resolution
	part from this line. There are higher resolution parts in the same
	family.) Is a minature TV camera using such a sensor with at least
	250x250 resolution  available yet?
	
	      Next best is a CCD line-transfer image sensor.  The
	better Pulnix units have these, and many robotic groups use them.
	What is the experience with these?  
	
	      The Sony Watchcam is a low-cost alternative.  Any
	experience here? 
	
	      What about TV transmitters and receivers?  I've seen a
	few TV Genie units around, but not only are they weak, they're
	illegal.  But they do show that such a transmitter need not be large,
	and there are bands in  which one can obtain appropriate licences.  I
	do need something about that size, though, say 4x2x2 or smaller.   Is
	there such a thing as FM TV gear, to improve the noise immunity? 
	
	      Has anyone dealt with the problem of camera
	stabilization and vibration isolation in a moving vehicle?  The
	Steadicam gyro approach seems overkill.	Sorbothane shock mounting is
	easy enough to do, but is it enough to get clear single frames?  Has
	anyone tried using data from accelerometers and rate gyros to
	stabilize an image electronically? 
	
	      Has anyone tried sending data back from a robot in the
	audio carrier of a TV signal or in the vertical retrace interval?  If
	so, with what hardware? 
	
	      Yes, I know it's a hard, ugly problem. 
	
						John Nagle
	

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

Date:     Fri, 27 May 88 10:00 H
From: <CHTEH%NUSEEV.BITNET@CORNELLC.CCS.CORNELL.EDU>
Subject:  ICIP'89 : 1989 IEEE International Conference on Image Processing



              1989 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING
                                    (ICIP'89)

                             PRELIMINARY ANNOUNCEMENT
                                       and
                              FIRST CALL FOR PAPERS


               The  1989  IEEE   International   Conference   on   Image
          Processing   (ICIP'89)  will  be  held  in  Singapore  on  5-8
          September, 1989.  The conference is jointly organised  by  the
          Computer  Chapter,  IEEE  Singapore  Section  and the National
          University of Singapore.  This will be the first international
          conference  solely devoted to image processing to be organised
          in this region.  The conference will include regular  sessions
          on  all  aspects  of  the  theory  and  applications  of image
          processing.   In  addition,  tutorials  by  eminent   speakers
          presenting  the  state-of-the-art  in  selected areas of image
          processing will be offered.

               Papers describing original work in all aspects  of  image
          processing  are invited.  Topics for regular sessions include,
          but are not limited to, the following :

             Image restoration/enhancement    Machine vision
             Video communications             Office image processing
             Image pattern recognition        Biomedical imaging
             AI vision techniques             VLSI implementation
             Remote sensing                   System architecture

               Authors are invited to submit four copies of an  extended
          summary of at least 1000 words of each of their papers to :

                    Technical Program Chairman, ICIP'89
                    c/o Meeting Planners
                    100 Beach Road, #33-01
                    Shaw Towers
                    Singapore 0718
                    Republic of Singapore

               The summary should contain sufficient detail, including a
          clear  description  of the salient concepts and novel features
          of the work.   The  summary  should  include  authors' names,
          addresses, affiliations, and telephone, telex and fax numbers.
          The authors should also indicate one  or  more  of  the  above
          topics that best describe the content of the paper.

               Proposals for  special  sessions  are  also  welcome  and
          should be addressed to the Technical Program Chairman.

          AUTHORS' SCHEDULE
               Submission of summary              9 January 1989
               Notification of acceptance         6 March 989
               Submission of final manuscripts    8 May 1989


For further information and questions, please either write to the Technical
Program Chairman at the above address OR E-mail to me at the following
E-mail addresses :
          eletehch@nusvm.bitnet
          chteh@nuseev.bitnet
For the period 28 May-18 June, 1988, please E-mail to me at the following
E-mail address :
          teh@ai.cs.wisc.edu  (ARPAnet at University of Wisconsin-Madison)

Dr. Cho-Huak TEH
Electrical Engineering Department
National University of Singapore
10, Kent Ridge Crescent
Singapore 0511
Republic of Singapore


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

Date:     Fri, 27 May 88 10:32 EDT
From: "Jacob N. George -- (313)947-1220" <"AESV01::GEORGE"@gmr.com>
Subject:  reply to request on DataCube
Subj:	reply to marc@acf8.nyu.edu

Your note hit on the major problems with Datacube:

1. Relatively difficult to program.
2. Very expensive - it usually requires many boards to perform any given
		    algorithm (avg oost is $3000 to $4000 per board).
3. Limited precision.

Of the three problems, #3 is the most serious.  Convolutions are limited to 
8 bit kernels.  In addition, the VFIR boards have only 8 bit inputs so it is 
not possible to concatenate boards and retain 16 bit output precision.
Only the MAX-SP and MAX-MUX boards are 16 bit input and output.  In addition,
none of the boards has a multiplier with greater than 8 bits.

Another difficulty I have encountered is limited memory.  To hold a 16 bit,
512 x 480 image requires 2/3 of a FRAMESTORE board.  This means that a
separate FRAMESTORE board is needed for each working ram that you use.

The vicom command set and fractional 16 bit data representation seem to be
better suited to general image processing.  I would guess that a vicom
system which resides on a SUN workstation is very price competetive with 
what would be required in a "comparable" datacube system.
 


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

Date: 3 Jun 88 08:35:55 GMT
From: andy@cs.hw.ac.uk (Andrew Wallace)
Subject: Quota Studentships in Computer Science
Organization: Computer Science, Heriot-Watt U., Scotland


                              Heriot Watt University

                          Department of Computer Science


        Applications are invited for 3 year research studentships  leading
        to   the  degree  of  Ph.D.,  commencing  in  October  1988.   The
        department has a wide range of research interests concentrated  in
        the areas of

                       Logic Programming and Databases
                       Expert Systems
                       Human-Computer Interaction
                       Computer Vision and Graphics

        Candidates must be British residents, and have or expect to obtain
        at  least an upper second class honours degree in Computer Science
        or a related subject, or an equivalent qualificaqtion.  They  must
        also display a strong motivation for research.  Enquires should be
        made in the  first  instance  to  Dr.  Andrew  Wallace,  Dept.  of
        Computer   Science,   Heriot-Watt   University,   79  Grassmarket,
        Edinburgh EH1 2HJ.


-Andrew Wallace			JANET : andy@cs.hw.ac.uk
				ARPA  : andy@uk.ac.hw.cs
				UUCP  : ..ukc!cs.hw.ac.uk!andy

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

Date: 8 Jun 88 23:54:54 GMT
From: munnari!dbrmelb.oz.au!ron%dbrmelb.dbrhi.OZ@uunet.UU.NET (Ron Sharpe)
Subject: Detection of pipe deterioration
Keywords: Image analysis, pipe probes & robots
Organization: CSIRO, Div. Construction and Engineering, Melb., Australia


This is a request for any research leads for a project to detect & analyse
defects inside buried pipes (non-metal) using visual and other sensors.
The object is to detect pipe cracking, water infiltration, corrosion,
roots, misalignment, wall strength and thickness and fat deposits.
We already have access to video images taken inside pipes down to
 150mm (6") diameter.

Any references dealing with image enhancement and AI analysis, use of sonics
and other detectors, probes that can operate inside pipes under
remote or self control would be greatly appreciated.

Dr Ron Sharpe, CSIRO Div of Building, Construction & Engineering,
PO Box 56, Highett, Vic 3190, Australia.
Ph (+613) 556 2211,  Fax (+613) 553 2819, Tlx AA 33766
e-mail: ron@dbrmelb.dbrhi.oz


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

Date: 10 Jun 88 08:05:07 GMT
From: munnari!cs.flinders.oz.au!mjb@uunet.UU.NET (Mike Brooks)
Subject: AI conference
Keywords: AI, conference, meeting, gurus
Organization: Flinders University of South Australia


          AUSTRALIAN JOINT ARTIFICIAL INTELLIGENCE CONFERENCE
AI'88 Secretariat, Department of Computer Science, The University of Adelaide,
         GPO Box 498, Adelaide, South Australia 5001, Australia.

                         15 - 18 NOVEMBER, 1988
                          ADELAIDE, AUSTRALIA

 The conference will be held from 15-18 November, 1988.  The three day period
 from 16-18 November will be given over to the main conference programme and
 Tuesday 15 November will be dedicated to a number of tutorial presentations
 by various speakers to fee paying groups by arrangement.

Keynote speakers for the conference will include Dr. William J. Clancey, Dr.
Scott Fahlman, and Dr M.P Georgeff.

Dr. Clancey received his Ph.D. degree from Stanford University in 1979.  His
doctoral dissertation described the acquisition of rule-based expertise
through dialogue.  He has been active in expert systems research since he
joined the MYCIN project in 1975.  His interests lie in computational
modeling of problem solving and the design of architectures for expert systems.
To investigate these issues, he has developed NEOMYCIN and instructional
programs based upon it.  He is presently Senior Scientist at the Institute
for Research on Learning at Palo Alto, California.

Dr. Fahlman received his Ph.D. degree from M.I.T. in 1977.  His doctoral
dissertation described the NETL architecture for representing large amounts
of real-world knowledge.  He is presently a Senior Research Computer Scientist
in the Computer Science Department of Carnegie-Mellon University.  His primary
research interest is in the area of massively parallel computing architectures
for A.I. problems.  Dr. Fahlman has also been very active in the development of
software tools for A.I. research.  He has also been a major participant in the
definition and standardization of the Common Lisp language.

Dr Georgeff is head of the new Institute for Artificial Intelligence in
Melbourne. He has spent many years researching at SRI in California and
will speak on real-time embedded systems.

Papers

Previously unpublished papers addressing any of the objectives of the
conference should be sent to the AI'88 Secretariat.  Persons in
industry are particularly encouraged to submit papers on their work.
Papers will be reviewed by a refereeing pannel and are due on 1st July
1988.  Authors will be given an opportunity to modify their papers.
Notification of acceptance will be on 31st August and final copy is
due on 30th September.

Registration Forms

Registration forms will be available from the AI'88 Secretariat at the above
address in early June 1988.


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

Date: 	Fri, 10 Jun 88 13:45:19 EDT
From: Venugopal Govindaraju <boulder!sunybcs!govind@HANDIES.UCAR.EDU>
Subject: source code for STEREOPSIS algorithms
Organization: SUNY/Buffalo Computer Science


I was wondering if I can borrow the source code for STEREOPSIS algorithms
which I could use to solve problems like finding correlation between
points in two different accumulator arrays.


I would really appreciate it if someone can e-mail it to me.

Thanks in advance

Venu
(govind@sunybcs.UUCP or govind@ellie.cs.buffalo.edu)

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

Date: Sat, 11 Jun 88 09:38:53 PDT
From: Vision-List-Request <vision@deimos.ads.com>
Subject: Multiple copies of Vision List Digest


Several subscribers have reported that they have received multiple copies of
the list.  If this is happening to you or your site, please mail me the
headers of all copies.  Some strange net bouncing is occuring.

Thanks,
Phil Kahn
Moderator Vision List Digest


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

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

