Copy Link
Add to Bookmark
Report
VISION-LIST Digest Volume 13 Issue 08
VISION-LIST Digest Tue Feb 22 11:26:42 PDT 94 Volume 13 : Issue 8
- ***** 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 FTP.TELEOS.COM
Today's Topics:
In lined fft
Multi-variate random number generator wanted
Program needed for LSHGC
Need Help!
Software for Polygon Detection
Digital camera sources?
High sensitivity video camera
TR: Extracting the Affine Transformation from Texture Moments
Error Reduction
----------------------------------------------------------------------
Date: Wed, 16 Feb 1994 10:43:00 +0000
From: dtm@ap.co.umist.ac.uk
Subject: In lined fft
Some time ago a heavily inlined fft function was advertised in Vision List.
Can anyone remember where the original came from? To prompt you, below is
the copyright/disclaimer from it.
Thanks
Tim Morris.
/* copyright 1990 Richard H. Krukar all rights reserved
Permission granted to buy, sell, or steal this software is granted.
The author retains the right to distribute this software freely, but
is not responsible for it's quality or maintainance. */
/* A split radix ( 2/4 ) FFT based on ( reference here ) This routine
is meant to be fairly efficient, but not exeptionally. Recursion is
used to sort out the butterflys. This should not matter because as
soon as the length reaches UNROL, everything is inlined elsewhere */
[ We still need to get a blazingly fast FFT into the Archive.
If you have it to share, please let me know.
phil... ]
------------------------------
Date: Thu, 17 Feb 94 10:55:00 +0100
From: paul.rosin@cen.jrc.it (Paul Rosin)
Subject: Multi-variate random number generator wanted
I want to add correlated Gaussian noise to a multi-band image. Although I've
written a program it's rather a hack and is really slow. Does anyone have any
better code for a multi-variate Gaussian random number generator?
Paul Rosin
Institute of Remote Sensing Applications
Joint Research Institute
Ispra (VA), 21020
Italy
email: paul.rosin@jrc.it
------------------------------
Date: Thu, 17 Feb 1994 09:13:30 -0500
From: James Davis <jdavis@cs.ucf.edu>
Subject: Program needed for LSHGC
Does anyone know where I can get a program (or code) that will
fit linear straight homogeneous generalized cylinders (LSHGC)
to a binary (0/255) image of a candy bar like object.
The object is completely solid (255), so continuous edges are
present.
Thanks.
James Davis
jdavis@cs.ucf.edu
------------------------------
Date: 17 Feb 1994 21:33:34 GMT
From: csp42@cc.keele.ac.uk (N.Zhu)
Subject: Need Help!
Hi! Everyone,
I want to get an artifical-neural-network approach to affine-
invariant curvature detection
How do visual systems recognize and discriminate curved contours
despite differences in position and direction of gaze of the observer?
One approach to this problem is to assume that curved contours are
given a representation that is viewpoint-invariant. To a good
approximation, changes in viewpoint may be represented by affine
transformations of the image (affine transformations preserve parallel
lines); from theoretical analyses, several affine-invariant
representations have been proposed as a basis for matching curved
contours.
What invariants can be obtained from training an artifical neural
network on families of affine-related curved lines?
Can you give some idea or some information about the above?
Thanks a lot!
Ning Zhu
Communication Department of
Keele University
Newcastle
Staffs ST5 5BG
U.K.
E-Mail:
csp42@seq1.cc.keele.ac.uk
------------------------------
Date: Tue, 15 Feb 1994 14:37:17 +0000 (GMT)
From: "W.Zhang" <W.Zhang@electronic-engineering.hull.ac.uk>
Subject: Software for Polygon Detection
Could anyone tell me where I can find software (via ftp) for 2D and 3D
arbitrary polygon detection?
Please reply me at the following address:
W.Zhang@electronic-engineering.hull.ac.uk
Thanks, any suggestions are welcome.
W. Zhang
------------------------------
Date: 21 Feb 1994 00:07:43 GMT
From: jam@grover.stat.washington.edu (John A. McDonald)
Organization: Univ of Wash , Stat Dept.
Subject: Digital camera sources?
Any pointers to sources for digital cameras will be greatly
appreciated. I am interested in both color and b&w, a range of
resolutions, and a range of prices. Any information about specs,
interfacing, prices, recommendations will also be greatly appreciated.
Thanks.
John Alan McDonald, Statistics, GN-22, U of W, Seattle, Wa. 98195
(206) 685 7438, 543 2517
jam@stat.washington.edu
------------------------------
Date : 17.Feb.94
From: Amir Schorr ami@math.tau.ac.il
Subject: High sensitivity video camera
Hi
We are looking for an high sensative B/W video camera , for an flourecent
microscopy project. The illumination is in the 10-100 mili Lux. and we consider
camera with or without integration whithin the camera. (propably by a shutter
control , for 0.25-1 sec integration, for the 0.001 Lux). The uper limit of 0.1 Lux can be seen by the HAD sony component.
Unfortunaly it is imposible to rely on the specification given by the companies.
We like to get information about video camera that work in these conditions,
and cost less than 3000 $ , based on personal expirience.
( For example , on Kohu : 4700, 4912, or Pulnix TM 9700)
Many Thanks in advance ;
Amir
------------------------------
Date: Thu, 17 Feb 1994 12:27:10 GMT
From: js2@eng.cam.ac.uk (J. Sato)
Organization: Cambridge University Engineering Department, UK
Subject: TR: Extracting the Affine Transformation from Texture Moments
The following technical report is available by anonymous ftp from the
archive of the Speech, Vision and Robotics Group at the Cambridge
University Engineering Department.
EXTRACTING THE AFFINE TRANSFORMATION
FROM TEXTURE MOMENTS
Jun Sato and Roberto Cipolla
Technical Report CUED/F-INFENG/TR167
Cambridge University Engineering Department
Trumpington Street
Cambridge CB2 1PZ
England
Abstract
In this paper we propose a novel, efficient and geometrically
intuitive method to compute the four components of an affine
transformation from the change in simple statistics of images of
texture. In particular we show how the changes in first, second and
third moments of edge orientation and changes in density are directly
related to the rotation (curl), scale (divergence) and deformation
components of an affine transformation. A simple implementation is
described which does not require point, edge or contour
correspondences to be established. It is tested on a wide range of
repetitive and non-repetitive visual textures which are neither
isotropic nor homogeneous. As a demonstration of the power of this
technique the estimated affine transforms are used as the first stage
in shape from texture and structure from motion applications.
**** How to obtain a copy ****
a) Via FTP:
unix> ftp svr-ftp.eng.cam.ac.uk
Name: anonymous
Password: (type your email address)
ftp> cd reports
ftp> binary
ftp> get sato_tr167.ps.Z
ftp> quit
unix> uncompress sato_tr167.ps.Z
unix> lpr sato_tr167.ps (or however you print PostScript)
b) Via postal mail:
Request a hardcopy from
Jun Sato,
Cambridge University Engineering Department,
Trumpington Street,
Cambridge CB2 1PZ,
England.
or email me: js2@eng.cam.ac.uk
------------------------------
Date: Thu, 17 Feb 1994 22:23:00 GMT
From: eledavis@ubvms.cc.buffalo.edu (Elliot Davis)
Organization: University at Buffalo
Subject: Error Reduction
Elliot Davis, Ph.D.
E-Mail: edavis@ubvms.cc.buffalo.edu
Phone: (716) 691-7235
I would greatly appreciate your thoughts on the:
ERROR TEMPLATE TECHNIQUE
The "Error Template" technique (patent 4,802,231) provides an
alternative method for reducing false alarms in pattern recognition
systems. In this approach, a pattern representing a mismatched
pattern is stored in the reference lexicon. It is a reference
pattern to an error rather then to what is desired. THIS IS DONE
WITH THE EXPECTATION THAT IF THE ERROR PATTERN OR A VARIATION OF IT
IS REPEATED IT WILL TEND TO BE CLOSER TO ITSELF THEN TO THE PATTERN
THAT IT FALSED OUT TO.
Preferential matching to an Error Template can result in the system
deciding that:
1. the test pattern was outside its reference vocabulary or
2. the test pattern falsely matched a desired reference pattern but
should be compared for possible matches to other desired reference
patterns and Error Templates. In this case the Error Template is
linked to specific desired reference patterns and is called a
Linked Error Template.
This technique should be tried on any pattern recognition system
that needs improvement in error reduction or speed. Pattern
recognition systems may be streamlined by reducing the amount of
desired reference templates while keeping the false alarm rate down
by the use of Error Templates. It is, relative to most other
techniques, extremely easy to implement. The Error Template can be
created and positioned in the comparison process in the same way as
a normal reference pattern. As all pattern recognition systems must
have a means of characterizing parameters to be stored and
compared, the applicability of the Error Template technique should
be widely and easily testable.
------------------------------
End of VISION-LIST digest 13.8
************************
