Copy Link
Add to Bookmark
Report
VISION-LIST Digest Volume 14 Issue 20
VISION-LIST Digest Thu Jun 01 09:58:23 PDT 95 Volume 14 : Issue 20
- ***** The Vision List host is TELEOS.COM *****
- Send submissions to Vision-List@TELEOS.COM
- Vision List Digest available via COMP.AI.VISION newsgroup
- If you don't have access to COMP.AI.VISION, request list
membership to Vision-List-Request@TELEOS.COM
- Access Vision List Archives via anonymous ftp to TELEOS.COM
Today's Topics:
Re: object-searching in an image
Real time image acquisition needed....
Hardware for Edge Det./Hough Trn.
Cheap Cameras!
help-frequency unit conversion
Deriche Source code!!
HELP needed: line detection in dot matrixes/images
Faces database
ANNOUNCEMENT: Public facial image database available by ftp
SoftInfo --Internet's Software Information Center
Photoshop and vision research?
Image understanding position
Position Available
PAMI Special Issue on Digital libraries: Representation and Retrieval
EANN 95 ftp,http sites
Call for Posters and Demos, ICCV Workshop
Image Processing Conference, Edinburgh, July 1995
Sun Annual Lecture, Department of Computer Science, Manchester University
Proceedings
----------------------------------------------------------------------
Date: Tue, 23 May 1995 16:42:17 -0500
From: "Raghunath K. Rao" <rrao@vision.iit.edu>
Subject: Re: object-searching in an image
Hi,
If you are searching for an object with known size and orientation,
then you can do better than normalized correlation. First of all
normalized correlation is very expensive in terms of computation,
because one has to normalize each correlation output (which is
easily computed via FFTs) with the local energy in the image.
Secondly, if you see the results, you'll find a lot of spurious
responses to irrelevant features in the image, and even the
correct peak will be quite broad and hard to detect.
In my PhD, I had developed a new method called Expansion matching (EXM),
which is based on decomposing the given image into basis functions
corresponding to shifted versions of the known template or object.
We also found an implementation that is much cheaper than
normalized correlation, and basically involves filtering the
given image with the EXM filter. All that is involved is a one-time
cost of designing the specific EXM filter for the object to
be detected. It is very robust to noise and clutter, and especially
robust to partial occlusion of the object. Not very sensitive
to minor changes of the size and orientation either.
What's nice is that the matching peak in the filter output
is VERY VERY sharp (ideally a delta function) compared to normalized
correlation, and one can easily detect the object. You can also
design the EXM filter to accomodate different levels of noise.
You can get details in the following references:
(omit the troff markups please :) )
.LI
Rao, K. R., and Ben-Arie, J.,
``Optimal Edge Detection Using Expansion Matching and Restoration,''
\fIIEEE Transactions on Pattern Analysis and Machine Intelligence\fR,
Vol. 13, No. 12,
December 1994, pp. 1169-1182.
.LI
Rao, K. R., and Ben-Arie, J.,
``Multiple Template Matching Using Expansion Matching,''
\fIIEEE Transactions of Circuits and Systems on Video Technology\fR,
Vol. 4, No. 5,
October 1994, pp. 490-503.
.LI
Rao, K. R., and Ben-Arie, J.,
``Non-Orthogonal Image Expansion Related to
Optimal Template Matching in Complex Images,''
\fIComputer Vision, Graphics, and Image Processing: Graphical
Models and Image Processing\fR,
Vol. 56, No. 2,
March 1994, pp. 149-160.
.LI
Ben-Arie, J., and Rao, K. R.,
``Optimal Template Matching by Non-Orthogonal Image Expansion Using Restoration,
''
\fIInternational Journal of Machine Vision and Applications\fR,
Vol. 7, No. 2,
March 1994, pp. 69-81
.LI
Ben-Arie, J., and Rao, K. R.,
``A Novel Approach For Template Matching by Non-Orthogonal Image Expansion,''
\fIIEEE Transactions on Circuits & Systems for Video Technology\fR,
Vol. 3, No. 1,
February 1993, pp. 71-84.
K. RAGHUNATH RAO /____/ \ /__/\ /______/\
email : rrao@chitra.ece.iit.edu | __ \ /| / _ \ \ / ___ \ |
sleepy (res) : (312)791-9428 | |__) |/ / /_\ \ \ | / | \ | |
sleepier (off) : (312)567-3407 | __ /\ | ___ | | | | | | | |
| | \ \ \ | / | \ | | | | |_| | |
It is the intonation and not | | |\ \ \ | | | | | | | \/__/ |/
the intention that matters!! |_|/ \_\/ |_|/ |_|/ \______/
------------------------------
Date: 30 May 1995 20:57:46 GMT
From: holder@central.cis.upenn.edu (Gaylord Holder)
Organization: University of Pennsylvania
Subject: Real time image acquisition needed....
Dear Colleagues:
We need to be able to acquire 20-30 secs of video, from
simultaneously 3 black and white CCD cameras at
rates of at least 10 frames per second. We do have the cameras,
we need frame grabber(s). The platform can be either a Sun or SGI.
Currently, we consider two options:
1. Use a color frame grabber and connect each b/w camera to each
channel
2. Explore the capabilities of Galileo.
I wonder if any of you has used a color frame grabber for this
purpose and with what results. Does any of you has vendor information?
Does any of you has a galileo video board?
Your help will be highly appreciated.
Gaylord
holder@central.cis.upenn.edu
------------------------------
Date: Fri, 26 May 95 00:37:33 EDT
From: murale@ee.eng.ohio-state.edu (Murale Kannapathipillai)
Subject: Hardware for Edge Det./Hough Trn.
Hi,
I have been working in a project in which a vision system is used to
navigate an autonomous vehicle. Sobel edge detection and hough
transform are jointly used in the approach. The problem we are facing
is the time it takes to run these processes. We are now thinking to
move this part of the software to a hardware based approach. Is
anyone familiar with special purpose hardware facilities either for edge
detection or for hough transform? I would greatly appreciate any
pointers/help/suggestions/references. My emil address is:
murale@ee.eng.ohio-state.edu
Murale K.
------------------------------
Date: Tue, 23 May 1995 10:47:19 -0500
From: Bruce Flinchbaugh <bef@csc.ti.com>
Subject: Cheap Cameras!
Regarding Chris Higginson's search for cheap cameras to use with PCs:
Connectix Corporation has a "QuickCam" for Windows camera that they say
will be available in August 1995 for $99. (Their camera for the Mac is
available now, and also sells for about $99.)
Info: interface to PC, CCD array, 6-bit monochrome video and stills, up to
320x240 pixels, 15 frames per second for 160x120 images (might be
limited by PC performance), fixed focus, 65 degree FOV, ...
Contact info:
Connectix Corporation
2655 Campus Drive
San Mateo, CA 94403
800-950-5880
info@connectix.com
Regards,
Bruce Flinchbaugh, bef@ti.com
Texas Instruments
------------------------------
Date: Wed, 31 May 1995 13:37:53 -0700
From: Krishna Allamneni <kmallamneni@ucdavis.edu>
Organization: University of California, Davis
Subject: help-frequency unit conversion
Hi,
I have a simple question regarding conversion from frequencies specified
in cycles per degree (cpd) (of visual angle) to normal image frequencies
(cps) (cycles per unit spacing).
For example, if I have a filter with cutoff 5.55 cpd, assuming that I am at
a viewing distance of d cm , for an image of size 256 what is the cutoff
frequency in terms of cps if the height of the picture is h?
what happens if the picture is now of size 512? what are the effects of
changing the viewing distance (now making d1=2d) as opposed to not
changing (d1=d)? what will be the cutoff in cps in the two cases?
the height of the picture is going to be 2h in both the cases.
what does this mean visually? can someone recommend a basic book/reference which
covers these things?
thank you in anticipation,
krishna
------------------------------
Date: Wed, 24 May 1995 16:57:38 +0200
From: cgarciab@etse.urv.es (Carlos Garcia-Barroso)
Subject: Deriche Source code!!
Dear world,
I would be so grateful if anyone could say me where can I find the
Deriche algorithm for edge detection (using separable recursive filters).
I am interested in C source code.
Tank you!
e-mail: cgarciab@etse.urv.es
Carlos Garcia-Barroso Villalonga
Dept. Enginyeria Informatica
Escola Tecnica Superior d'Enginyeria
Universitat Rovira i Virgili
Autovia de Salou s/n
43006. Tarragona. SPAIN.
Tlf. 34 77 55 96 79
Fax. 34 77 55 97 10
------------------------------
Date: 29 May 1995 13:31:11 GMT
From: rolland@laforia.ibp.fr (ROLLAND Pierre-Yves Equipe de J.-G. Ganascia)
Organization: University Pierre & Marie Curie PARIS6, CNRS, LABORATORY LAFORIA
Subject: HELP needed: line detection in dot matrixes/images
Keywords: image analysis, dot matrix, line detection, edge linking
Hello:
I am a researcher in Artificial Intelligence currently looking for
existing computer programs (or algorithms) able to automatically
detect lines or approximate-lines in matrixes; more precisely:
In the simplest situation, I deal with numerical 2-D (classical)
boolean matrixes, i.e. with elements 1 or 0, such as dot matrixes
obtained from aminoacid sequence in molecular biology. The problem is
to detect line segments typically 5-6 elements-long as on fig. 1
below, or discontinuous/approximate segments as on fig.2:
oooooooooooooooo oooooooooooooooo
ooooo1oooooooooo oo1ooooooooooooo
o1ooooo1oooooooo ooo1oooooooooooo
oo1oooooo1oooooo oooo1ooooooooooo
ooo1ooooooo1oooo oooooooooooooooo
oooo1oooooooo1oo ooooooooo1oooooo
ooooo1ooooooooo1 oooooooooo1ooooo
oooooooooooooooo ooooooooooo1oooo
Fig. 1 Fig.2
In more complicated ones, I deal with *multidimensional*-,
symbolic-elements- matrixes. Here also I am trying to detect
alignments (pseudo-linear clusters) of elements having equal attribute
values (e.g. elements all having a "color-attribute" value equal to
"red" and a "shape-attribute" value equal to "triangle").
A brief literature search in the fields of either computer vision,
pattern recognition and molecular biology showed that a number
programs/algorithms accomplishing this type of tasks exist, but it's
hard to have a clear view so as to focus on the right one(s) for my
particular problem.
Anyone's help will be greatly appreciated. Thanks.
Pierre-Yves ROLLAND
LAFORIA/IBP/Universite Paris 6
4 place Jussieu
75005 Paris, FRANCE
e-mail: rolland@laforia.ibp.fr
------------------------------
Date: 17 May 1995 18:32:30 GMT
From: mcarreir@ls.fi.upm.es (Miguel Carreira Perpignan)
Organization: Facultad de Informatica - (UPM)
Subject: ANNOUNCEMENT: Public facial image database available by ftp
Hello,
The USENIX faces archive is a public database, accessible by ftp, that
can be of use to people working in the fields of human face recognition,
classification and the like. It currently contains 5592 different faces
(taken at USENIX conferences) and is updated twice each year. The images
are mostly 96x128 greyscale frontal images and are stored in ascii files
in a way that makes it easy to convert them to any usual graphic format
(GIF, PCX, PBM etc.). Source code for viewers, filters, etc. is
provided. Each image file takes approximately 25K.
You can find this database in src.doc.ic.ac.uk:/pub/packages/faces.
Please have a look at the README file prior to accessing the archive in
order to find your way in it, and do NOT do a directory listing in the
top directory of the face archive, as it contains over 2547 entries.
According to the archive administrator, Barbara L. Dijker
(barb.dijker@labyrinth.com), there is no restriction to use them.
However, the image files are stored in separate directories
corresponding to the Internet site to which the person represented in
the image belongs, with each directory containing a small number of
images (two in the average). This makes it difficult to retrieve by ftp
even a small part of the database, as you have to get each one
individually.
A solution, as Barbara proposed me, would be to compress the whole set
of images (in separate files of, say, 100 images) and maintain them as a
specific archive for research on face processing, similar to the ones
that already exist for fingerprints and others. The whole compressed
database would take some 30 megabytes of disk space. I encourage anyone
willing to host this database in his/her site, available for anonymous
ftp, to contact her for details (unfortunately I don't have the
resources to set up such a site).
Please consider that UUNET has graciously provided the ftp server for
the FaceSaver archive and may discontinue that service if it becomes a
burden. This means that people should not download more than maybe 10
faces at a time from uunet.
A last remark: each file represents a different person (except for
isolated cases). This makes the database quite unsuitable for training
neural networks, since for proper generalisation several instances of
the same subject are required. However, it is still useful for use as
testing set on a trained network.
Included is also a version of the archive with the images in 48x48 XBM
format (black and white) that takes 2.6M (15M unpacked), but they
probably contain too little information for automatic recognition.
Miguel A. Carreira-Perpinan
Dept. Computer Languages and Systems (DLSIIS)
Technical University of Madrid
mcarreir@moises.ls.fi.upm.es
PS. Please bear in mind that I am in no way related with the USENIX
faces archive. I simply found it by chance and --after having checked
with the archive admin-- posted this article for general knowledge.
------------------------------
Date: Tue, 30 May 1995 19:54:15 GMT
From: Dennis Hamilton <dlhamilton@icp.com>
Organization: IQuest Network Services
Subject: SoftInfo --Internet's Software Information Center
[ Of interest as a way to find relevant commercial CV/IP software.
phil... ]
RE: SoftInfo -- http://www.icp.com/softinfo/
I wanted to share with your newsgroup the announcement that
SoftInfo, the world's most extensive data base of information
on software products and suppliers, now is available free of
charge on the World Wide Web. SoftInfo, the ICP Software
Information Center, contains complete details on more than
16,500 software products, tools, languages, and utilities for
everything from laptops to supercomputers. It also contains
profiles of, and direct contact information for, more than
4,500 software suppliers. If the suppliers have home pages of
their own, you can link to them right from SoftInfo.
The mission behind SoftInfo is to give seekers of software
information a single-address gateway to virtually every player
and product in software.
SoftInfo is produced and maintained by ICP, Inc, the world's
oldest software information company. Founded in 1966, ICP has
published software information on more than 150,000 products
for over one million users. Media have included magazines,
directories, newsletters, floppy disks, CD-ROM, and now on the
WWW. On SoftInfo, users can search by product name, company
name, full-text key words, platforms, or by alpha browsing.
We'd very much appreciate your help in getting out the word
about SoftInfo. In turn, let me know when you have news of
interest, and we'll talk about it in our news pages.
------------------------------
Date: 1 Jun 1995 11:27:59 -0500
From: jorn@mcs.com (Jorn Barger)
Organization: The Responsible Party
Subject: Photoshop and vision research?
Summary: are they converging yet?
Is there a program like Photoshop, yet, that encourages you to 'explain'
the objects within an image-- marking out regions that you tag as
wall, roof, sky, cloud, treetrunk, face, clothing, grass, concrete,
etc?
...So that it can intelligently suggest/ assist with editing
operations???
thanks for any indulgences...
j
jorn@mcs.com
------------------------------
Date: Tue, 30 May 95 18:18:09 MDT
From: chd@aragtap.den.mmc.com (Chris Debrunner)
Subject: Image understanding position
Our image understanding group expects to have several positions for
image understanding/computer vision researchers and programmers
opening up over the next six months. We work in the areas of
automatic target recognition and automated exploitation aids for image
analysis. We are looking for researchers with BS, MS and PhD degrees
(or comparable experience). The skills of interest to us are (in
rough order of importance):
Image Understanding / Computer Vision
C/C++ Programming
Object Oriented Design
Parallel Algorithms and Programming
Khoros
UNIX
LISP Programming
Knowledge-Based Systems/Artificial Intelligence
Relational Database Programming and Administration
A solid background in math and physics is also desirable.
Applicants selected may be subject to a security investigation for
access to classified information.
For more information, please send email to Chris Debrunner at
chd@aragtap.den.mmc.com. Resumes may be sent to the same email
address or faxed to (303) 977-7946.
------------------------------
Date: Thu, 1 Jun 95 15:54:22 +0200
From: pitas@zeus.csd.auth.gr (Ioannis Pitas)
Subject: Position Available
RESEARCH POSITION
Image sequence processing/analysis/coding
The Department of Informatics, University of Thes-
saloniki has an open research position for 12 months (maximum),
starting 1/6/95. Its topic is image sequence
processing/analysis/coding. The position is financed by an EC
research project. Prospective applicants must be citizens of
EC/EFTA countries (excluding Greece). The position is at postdoc
level (preferably). However, good candidates that do not have a
PhD degree could be accepted as well. A PhD degree or proven
research experience in relevant topics (digital image
processing), very good knowledge of English and strong interest
in academic research are highly desired.
A indicative list of possible research topics would be the fol-
lowing:
1) Image sequence filtering (nonlinear, motion-compensated tech-
niques etc).
2) Color image processing/analysis (nonlinear, use of various
color spaces).
3) Motion field processing (smoothing, moving object segmenta-
tion)
4) Nonlinear techniques in image sequence coding (e.g. use of
morphological techniques in model-based coding).
Work on all these topics is currently performed at our group. The
exact topic of the new researcher will match his/her experience
aiming at maximum productivity.
Prospective applicants can send their resume and reference let-
ters to
IOANNIS PITAS
DEPARTMENT OF INFORMATICS
UNIVERSITY OF THESSALONIKI
THESSALONIKI 54006, PO BOX 451
GREECE
TEL +30-31-996304
FAX +30-31-996304
EMAIL pitas@zeus.cd.auth.gr
Ioannis Pitas
Professor, Department of Informatics
University of Thessaloniki, Thessaloniki 54006, PO Box 451
GREECE
Tel. +30-31-996304
Tel. +30-31-996361 (lab)
Fax: +30-31-996304
email: pitas@zeus.csd.auth.gr
http://poseidon.csd.auth.gr/
old email (still working): pitas@vergina.eng.auth.gr
------------------------------
Date: Wed, 31 May 95 12:30:09 -0400
From: Roz W. Picard <picard@media.mit.edu>
Subject: PAMI Special Issue on Digital libraries: Representation and Retrieval
FINAL CALL FOR PAPERS
Special Issue of the IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE
INTELLIGENCE on Digital libraries: Representation and Retrieval
Papers are solicited for a special issue of the IEEE TRANSACTIONS ON
PATTERN ANALYSIS AND MACHINE INTELLIGENCE which will address the
subject of Digital libraries: Representation and Retrieval. The Guest
Editors for the special issue will be Rosalind W. Picard and Alex P.
Pentland, both of the M.I.T. Media Laboratory, USA.
With the rapid increase in worldwide networks, people will soon have
access to massive libraries of digital text, sound, image, video and
other special-purpose data. The biggest obstacle to this
``information glut'' is that the technology for organizing, querying,
and retrieving data from these multimedia libraries is still in its
infancy.
Research in pattern analysis and machine intelligence is needed to
provide tools for accessing the content of these databases. The
papers in this special issue will describing novel and significant
tools and techniques that facilitate access to the content of large
data collections. Manuscripts will not be accepted if they have been
previously published, or if they describe algorithms that have not
been carefully evaluated on large databases. Possible topics for
papers submitted to the special issue include, but are not limited to:
* Robust recognition and retrieval of classes of information, such as
``find all photos containing trees'' or ``all places where there is
laughter.''
* Search on compressed representations, and new representations that
facilitate efficient semantic or perceptual searching.
* Intelligent searching through huge amounts of text or data, e.g. for
DNA matching or drug design.
* Automatic annotation to generate descriptions of data, especially
based on combinations of visual, acoustic, and motion (temporal)
features.
* Recognition of video content invariant to viewing conditions, e.g.,
``Find all other shots of this scene.''
* Computational measures of perceptual similarity: especially texture,
shape, color, that have been evaluated on a large set of data, e.g.,
``computer, are there any other drum patterns that sound like this?''
or ``find fabrics which look most like this one.''
* Segmentation of video/soundtrack/image as applied to database search
and retrieval.
* Automatic extraction of video keyframes, or analogous summarizing
information in non-visual media.
All papers will be reviewed by the guidelines of the transactions.
Please submit four copies of your paper to:
Prof. Rosalind W. Picard
MIT Media Laboratory, E15-392
20 Ames Street
Cambridge, MA 02139
Schedule:
Deadline for submission of manuscripts: June 30, 1995
First set of reviews to authors: December 1, 1995
Final manuscripts due: March 1, 1996
Publication of special issue: November, 1996
For further information, contact
Rosalind W. Picard (picard@media.mit.edu)
or
Alex P. Pentland (sandy@media.mit.edu).
------------------------------
Date: Sat, 27 May 1995 16:23:40 +0300
From: EANN-95 Konferensomrede VT <eann95@ra.abo.fi>
Subject: EANN 95 ftp,http sites
The EANN '95 conference program is in /pub/vt/ab/eann95schedule.ps.Z
and can be picked up by anonymous ftp from ftp.abo.fi. EANN '95 home
page is at http://www.abo.fi/~abulsari/EANN95.html.
EANN '95 organisers
------------------------------
Date: Thu, 25 May 95 17:11:19 -0400
From: Laurie Pillsbury <lawp@media.mit.edu>
Subject: Call for Posters and Demos, ICCV Workshop
CALL FOR POSTERS AND DEMONSTRATIONS
In conjunction with the ICCV Workshop on Visual Information
Management, June 19, Bartos Theater, The Media Laboratory, M.I.T.
In the last few years, many researchers have begun building systems
for visual information management. Such systems encompass central
aspects of databases, pattern recognition, image understanding,
interface design, and AI reasoning systems. The challenge for vision
researchers is to develop tools for analyzing the semantic content of
video and image, representing that content in a way that can be
efficiently searched and compared, and finally delivering it to users
in the form of useful, focused presentations.
To help stimulate research and communication in this area, we are
organizing a workshop on visual information management, sponsored by
NSF and ARPA. This one-day workshop will consist of presentations and
discussions involving researchers from around the world.
As part of this workshop we ask interested researchers to submit
proposals for poster presentations and computer demonstrations.
Poster presentations should be no more than 500 words in length,
including title, author list, email addresses, postal addresses, and a
summary description of the poster's content. The submission should be
by email only, and must arrive before May 31, 1995. Notification of
acceptance will be by June 7, 1995.
Proposals for computer demonstrations should be no more than 500 words
in length, including title, author list, email addresses, postal
addresses, and a summary description of the demonstration's content
and a detailed description of required space, time, and computation
resources. The submission should be by email only, and must arrive
before May 31, 1995. Notification of acceptance will be by June 7,
1995.
Submissions and questions should be sent to either Alex Pentland
(sandy@media.mit.edu) or Rosalind Picard (picard@media.mit.edu).
------------------------------
Date: Tue, 23 May 95 09:48:00 PDT
From: "Griffiths, Sheila" <SGriffiths@iee.org.uk>
Subject: Image Processing Conference, Edinburgh, July 1995
International Conference on
'Image Processing and its Applications'
Heriot-Watt University, Edinburgh, UK: 3 - 6 July 1995
This conference is the fifth in the successful series on image processing
organised by the Institution of Electrical Engineers and aims to provide a
forum for the exchange of new results in the area of image processing. The
conference will consist of 3 days of parallel sessions and will be preceeded
by a day of tutorials, details of which follows. If you would like to
register for the tutorial or would like to receive a copy of the full
conference programme please contact:
IPA95 Secretariat
IEE Conference Dept
Savoy Place, London WC2R OBL
Tel: + 44 (0)171 344 5477/8
Fax: + 44 (0)171 240 8830
email conference@iee.org.uk (mention IPA95 in message)
TUTORIAL PROGRAMME AND REGISTRATION
MONDAY, 3 JULY 1995
Session 1 09.00 hrs
Introduction to Concepts of Mathematical Morphology
Lecturer: S Marshall
University of Strathclyde
Scotland
This tutorial lecture will cover the basic concepts of morphology, erosion
dilation, opening
and closing. The following aspects of mathematical morphology will be
described: hit or
miss transform; rolling ball; properties of morphology extensivisty,
increasing,
idempodence; connectivity, skeletons; multiscale problems; pyramids and
pattern
spectrum; grey scale morphology. Reference will be made to applications
involving
medical image processing, segmentation and multimodality data fusion.
10.30 hrs Coffee Break
Session 2 11.00 hrs
Neural Networks for Image Processing Applications
Lecturer: G Qiu
University of Derby
England
This tutorial lecture will present an overview of some neural network image
processing
applications. The following will be presented:
o An overview of three neural network models: Hopfield network,
Backpropagation network and the Kohonen network.
o A block truncation coding technique based on Hopfield neural networks.
o Image compression based on backpropagation networks.
o Self-organising visual pattern generation with application to image
coding.
12.30 hrs Lunch Break
Session 3 13.30 hrs
Approaches to Image Segmentation
Lecturer: D Telfer
University of Central Lancashire
England
The topics to be covered by this tutorial lecture are:
Segmentation by intensity: binary and grey-level images. Edge detection.
Region
growing. Feature detection and clustering. Texture-based segmentation.
Digital and
optical Fourier filtering.
15.00 hrs Tea Break
Session 4 15.30 hrs
Image Processing and Restoration
Lecturers: D M Holburn and W O Saxton
Cambridge University, UK
This tutorial lecture will cover the use of digital computers in the
acquisition,
enhancement, analysis and general manipulation of image data, and concepts
will be
illustrated using a case study involving scanning electron microscopy. This
will also
include measurement and 3-D reconstruction.
17.00 hrs Comfort Break
Session 5 17.15 hrs
ITU and ISO Standards for Video Compression Coding
Lecturer: G Morrison
Manager, Video Processing
BT Laboratories, UK
This tutorial lecture will introduce the ITU and ISO standards which are key
enablers for
the economic inclusion of pictures in mass market storage and transmission
applications.
The basic principles behind them will be explained first, followed by more
detailed
illustrations of the specific standards intended to meet different
requirements and
operating conditions.
18.45 hrs End of Tutorial Programme
Tutorial Registration
The fee for the tutorial sessions is L 90.00 (including L13.40 VAT). This
includes lunch,
morning and afternoon refreshments and a set of tutorial notes. Those
wishing to
participate should complete and return the registration form.
Tutorial notes will be available for collection from the secretariat between
08.15 and 09.00 hrs prior to the
first session.
------------------------------
Date: 1 Jun 1995 15:10:11 GMT
From: timc@cs.man.ac.uk (Tim Clement)
Organization: Dept of Computer Science, University of Manchester, U.K.
Subject: Sun Annual Lecture, Department of Computer Science, Manchester University
1995 Sun Annual Lecture in Computer Science
University of Manchester
August 31st -- September 1st
Teaching Vehicles to See:
Dynamic Machine Perception for Motion Control
Ernst D. Dickmanns
Professor Ernst D. Dickmanns will give a series of eight lectures on
dynamic machine perception for motion control. The lectures will be
illustrated with video of the self guiding vehicles that his group has
developed in action.
Dr.-Ing. E.D. Dickmanns is Professor for Control Engineering at the
Aerospace Technology Department of the Universitat der Bundeswehr, Munich.
Registration for the lecture series is 90 pounds (45 pounds for
registered students), which includes lunch each day, and supporting
material for the lectures.
Further information, including a synopsis of the lectures and
registration forms, is available on WWW at
http://www.cs.man.ac.uk/events/sun-lecture.html
or by e-mail request to annual-lecture@cs.man.ac.uk.
------------------------------
Date: Wed, 31 May 1995 08:47:50 +0200
From: Roger Mohr <Roger.Mohr@imag.fr>
Subject: Proceedings
The Workshop on
Geometrical Modeling and Invariants for Computer Vision
was held in Xi'an, China, the 27, 28, 30 April 1995. Exciting new developments
were presented and the text of the communications published by Xidian
University Press, Xi'an (ISBN 7-5606-0383-1).
There are 60 papers included in the proceedings. The papers are edited
in five sections: Curved 3D Objects, Image Geometry and Algebra, 3D
Projective Invariants, 3D Reconstruction and Robots, and Shape
Invariants (see below for the table of content). The proceedings
consists of 404 pages with a softcover.
Proceedings can be obtained by ordering (only this year) from
Daniele Herzog,
Lifia-Inria,
46 av. Felix Viallet
38031 Grenoble, France
fax: (33) 76 57 46 02 tel: (33) 76 57 48 06
email : daniele.herzog@imag.fr
Order form :
number of GMICV proceedings : n =
Price (including shipping in Europe) : n x $45 or 225FF
out of Europe : : n x $50 or 250FF
banking expenses (if applicable see below) :
total amount :
Payment can be made in two ways:
1. Free of charge : By money transfer to:
Institut National Polytechnique, Agent comptable
bank bank code place code account control key
TPGrenoble, 100071 38000 00003000141 43
the money transfer should state: "for UC M501 X2"
(this is just the code concerning these proceedings)
2. By sending a cheque to Daniele out to
Agent comptable de l'INPG.
In this case add $8.40 (or 42FF) for banking expenses
(except for cheques coming from a French bank).
No cash, no credit card, sorry.
Table of contents
=================
KEYNOTE SPEECH 1
Geometric Invariants for Computer Vision
LEE Chung-Nim
SESSION 1: CURVED 3D OBJECTS
O1-1 Affine Length and Affine Dimension of a 1-set R2
Dibos F.
O1-2 Scale-spaces and Affine Curvature
Faugeras O., Keriven R.
O1-3 Modeling 3D Objects with Patches of Quadratic Surfaces: Application
to the Recognition and Locating of Anatomic Structures
Bricault I., Monga O.
O1-4 Derivative Computation by Multiscale Filters
Ma S.D., Li B.C.
O1-5 Delaunay Triangulation of Flexible Contour Model
Li W., Zheng N.N.,Y uan L.X.
O1-6 Segmentation and Localization of 3D Objects Using Implicit Polynomials
Kaveti S., Teoh E.K., Wang H.
O1-7 3D Symmetry Detection Using Extended Gaussian Image
Sun C.M., Sherrah J.R.
O1-8 Representing Generalized Cylinders
Naeve A., Eklundh J.O.
O1-9 B-spline Contour Fitting and Transform Representation for Computer
Vision
He P.L., Wang Y.P., Wang T.Y., Liang Y.J.
O1-10 The Statistical and Syntactical Modeling and Recognition of Flexible
Objects
Zhu S.C., Yuille A.L.
SESSION 2: IMAGE GEOMETRY AND ALGEBRA
O2-1 Fuzzy Solid Sets and Morphological Image Algebra
Liu L.R.
O2-2 The Study of 3D-from-2D Using Elimination
Werman M., Shashua A.
O2-3 On the Geometry and Algebra of the Point and Line Correspondence
between N Images
Faugeras O., Mourrain B.
O2-4 Trilinearity and the 3D-from-2D Reconstruction Problem
Shashua A., Werman M.
O2-5 Motion Clustering Using the Trilinear Constraint over Three Views
Torr P.H.S., Zisserman A., Murray D.W.
O2-6 A Matching Method Based on Hausdorff Distance
Wu Y., Ding M.Y., Peng J.X.
O2-7 Geometrical Invariant for Computer Vision Based on Mahalanobis Distance
Xuan G.R., Chai P.Q.
O2-8 Articulation Detection for Locally Rigid Objects
Sinclair D.
O2-9 Object Modelling and Motion Analysis Using Clifford Algebra
Bayro-Corrochano E., Joan L.
SESSION 3: 3D PROJECTIVE INVARIANTS
O3-1 Computing Three-dimensional Projective Invariants from a Pair of Images Using the Grassmann-Cayley Algebra
Csurka G., Faugeras O.
O3-2 Application of the Twisted Cubic to Model Based Vision
Maybank S.J.
O3-3 Geometric Invariance Used in Stereo Vision for Model Building
Lin X.Y., Deng W.
O3-4 Joint Projective Invariants for Five Coplanar Lines
Xu Z.W., Wu C.K.
O3-5 Grouping and Invariants using Planar Homologies
Gool L.V., Proesmans M., Zisserman A.
O3-6 Invariant of a Pair of Non-coplanar Conics in Space
Quan L.
SESSION 4: 3D RECONSTRUCTION AND ROBOTS
O4-1 A Global Stereo Vision Method Based on Wu-solver
Xu C.X., Shi Q.Y., Cheng M.D.
O4-2 The Advantage of Mounting a Camera onto a Robot Arm
Horaud R.,Mohr R., Dornaika F., Boufama B.
O4-3 Affine Calibration of Mobile Vehicles
Beardsley P., Zisserman A.
O4-4 Automatic and Accurate Object Positioning using Targets
Boufama B., Mohr R., Morin L.
O4-5 Stereo Correspondence for Planar Curves Based on Their Invariant
Wen W., Yuan B.Z.
O4-6 Interpreting Axonometric Drawing Based on Lines
Gao M.T., Qu S.R.
O4-7 Different Paths towards Projective Reconstruction
Rothwell C., Faugeras O., Csurka G.
O4-8 Calibrating a Binocular Stereo through Projective Reconstruction Using
Both a Calibration Object and the Environment
Zhang Z.Y., Faugeras O., Deriche R.
O4-9 A New Method of 3D Objects Reconstruction from Range Data
Tian J.,Dai R.W.
O4-10 Object Modeling via Sparse Range Images
Zeng J.C., Xu G.Y.
KEYNOTE SPEECH 2
Generic Object Recognition and Quasi-invariance
Gerard MEDIONI
SESSION 5: SHAPE INVARIANTS
O5-1 Development of 3D Invariants Using Linear Algebra and Tensor Theory
Burel G., Henocq H., Catros J.Y.
O5-2 PRSI Shape Classification Using Radius Vectors
Ye X.Y., Qi F.H.
O5-3 Recognition of Planar Objects over Complex Backgrounds Using Line
Invariants and Relevance Measures
Startchik S., Rauber C., Pun T.
O5-4 A Projective Invariant Metric for Measurement of Similarity Between
Two Polygons
Batatia H.
O5-5 3-D Object Description for Recognition
Wang R.S., Liu F.
O5-6 On the Extraction of the Face Features
Li J.G., Liu C.Y., Qi Z.Y.
O5-7 On Modelling, Extraction, Detection and Classification of Deformable
Contours from Noisy Images
Lai K.F., Chin R.T.
POSTER SESSION
P-1 Invariants of Fourier Descriptor and its Relation with Moments
Invariants
Ma S.D., Li B.C.
P-2 Research on Using Moment Invariants in Scene Matching
Wu Y., Ding M.Y., Peng J.X.
P-3 Camera Auto-calibration from Known Motion
McLauchlan P.F.
P-4 Interreflections Are Useful Rather Than Harmful in Shape Recovery of
a Concave Polyhedron from a Single Image
Yang J., Ohnishi N., Sugie N.
P-5 Modeling Facial Image with Flexible Contour Method
Li W., Zheng N.N., Yuan L.X.
P-6 Affine Normalization of Planar Regions by Moments Using a New
Separation Method
Voss K., Suesse H., Rothe I.
P-7 Extraction of Corner-Edge-Surface Structure from Range Images Using
Mathematical Morphology
Chen C.S., Hung Y.P., Wu J.L.
P-8 CV/CAD Based 3D Object Modeling System
Deng S.W., Yuan B.Z.
P-9 An Optoelectronic Approach of Calculating Morphological Pattern
Spectrum and its Application on Pattern Recognition
Liang F., Liu L.R., Wang B.Q., Peng H.F.
P-10 Finding the Center: Using Incidence to Recover Geometric Features
from Single, Monocular Views
Coe D.H., Fallon J.B., West R.L., Abbott A.L.
P-11 Obtaining Correspondences from 2D Perspective Views with Wide Angular
Separation of Non-coplanar Points
Georgis N., Petrou M., Kittler J.
P-12 Adaptative Filtering and Geometrical Invariants in Face's Depth Maps
Prinet V.,Monga O.
P-13 A Method of Invariant Image Processing
Tkacheva O.
P-14 Adaptive Contour Detection (ACD) with Two-Dimensional Continuous
Wavelet-Transform
Li G.
P-15 Perspective Invariance Segmentation of Planar Curves
Xu Z.W., Wu C.K.
P-16 Towards a Reliable Extraction of Euclidean Differential Invariants In
3D Medical Images
Lengagne R., Monga O., Cong G., Ma S.D.
------------------------------
End of VISION-LIST digest 14.20
************************
