Need literature! Subject: Object recognition in the human visual system

Steve Lehar slehar at alewife.bu.edu
Thu Oct 17 08:38:28 EST 1991


> Is there anyone who has (or knows about) literature to the  subject of
> object recognition in the human visual system ?  The literature I have
> is (or is derived  of)  the work of  Hubel  and  Wiese on the  striate
> cortex (hypercolumn theory).

Perhaps this is not exactly what you are looking for, but I believe it
is the next logical step upwards from the Hubel and Wiesel data.  What
Hubel  and  Wiesel showed was  that   the  basic "language" or   "data
representation"  of the  visual  system  is  moving or static oriented
edges (as  opposed to pixels,  in computer image representation).  The
question    is, what   happens  next  in    visual  processing?   What
"computations" are performed on this raw information?

Grossberg  & Mingolla have  answered this  question using a  different
source of information, that being perceptual data, visual illusions in
particular.  Why do we  perceive certain things  that are  clearly not
there in the input image?  Grossberg & Mingolla  have shown how we can
use  such perceptual  data to infer  certain processing  mechanisms of
biological vision, and have expressed those  mechanisms in the form of
dynamic neural models which can be simulated by  computer in  order to
reproduce  the original illusions.   Using this   method Grossberg and
Mingolla have been able to reproduce and explain  a very wide range of
diverse visual illusions.

What  is  interesting about their model  is that it reveals  a complex
mechanism with very interesting properties, with a level of robustness
and  adaptability quite different from anything  conceived by  man for
computer   image  processing.    Essentially, the  model  consists  of
multiple cooperative  and  competitive  interactions   in a  recurrent
multi-level hierarchy that optimizes the multiple constraints posed by
the image by  a simultaneous   relaxation in  all  the  levels of  the
hierarchy.

Now  this  can by  no  means  be  described   as  "high  level" object
recognition, but rather, it represents  a disambiguation of the lowest
level input, or "image  enhancement".  The nature of this  enhancement
however affords an insight into the way nature solves visual problems,
and  gives us a hint as  to how  such higher level processing might be
accomplished  in  the    visual   system.  In   particular, this model
emphasizes the  importance of  recurrent  and long  range interactions
between  local computational   elements  in  order to  achieve  global
consistancy throughout the system.  My own work, which is an extension
to  the Grossberg and Mingolla model  emphasizes this aspect of  their
model.

I have prepared a  lengthy text  file that outlines the  prinpal ideas
behind these interesting  models, which I  would be happy  to email to
you, or other interested parties, on request.


Stephen  Grossberg  &   Ennio Mingolla NEURAL  DYNAMICS  OF PERCEPTUAL
GROUPING: TEXTURES, BOUNDARIES AND EMERGENT SEGMENTATIONS Perception &
Psychophysics (1985), 38 (2), 141-171.   This work presents the  BCS /
FCS model   with  detailed psychophysical   motivation   for the model
components and computer simulation of the BCS.

Stephen  Grossberg  & Ennio   Mingolla   NEURAL   DYNAMICS OF  SURFACE
PERCEPTION: BOUNDARY WEBS,   ILLUMINANTS,     AND  SHAPE-FROM-SHADING.
Computer Vision,  Graphics  and Image  Processing (1987) 37,  116-165.
This  model  extends the BCS  to explore  its response to gradients of
illumination.  It is mentioned here because of an elegant modification
of the second competitive stage that was utilized in our simulations.

Stephen Grossberg &  Dejan Todorovic NEURAL  DYNAMICS  OF 1-D AND  2-D
BRIGHTNESS PERCEPTION Perception and Psychophysics (1988) 43, 241-277.
A beautifully  lucid summary  of BCS /  FCS  modules  with 1-D and 2-D
computer  simulations  with  excellent  graphics  reproducing  several
brightness   perception illusions.    This algorithm   dispenses  with
boundary completion,  but in return  it  simulates  the FCS operation.
Reprinted  in  NEURAL  NETWORKS   AND  NATURAL  INTELLIGENCE,  Stephen
Grossberg Editor, MIT Press (1988) Chapter 3.

Lehar S., Worth A. MULTIPLE RESONANT BOUNDARY CONTOUR SYSTEM.  In:
PROGRESS IN NEURAL NETWORKS volume 3 (Ed. by Ablex Publishing Corp.)
In print.
	 
Lehar S.,  Worth   A. MULTIPLE  RESONANT    BOUNDARY  CONTOUR  SYSTEM.
Proceedings of the SPIE, Vol. 1469  (Applications of Artificial Neural
Networks Conference) April 1991.

Lehar S.,  Worth A. MULTI   RESONANT BOUNDARY CONTOUR  SYSTEM,  Boston
University,   Center   for  Adaptive   Systems   technical  report   #
CAS/CNS-TR-91-017

  Boston University
  Center for Adaptive Systems
  111 Cummington Street, Second Floor
  Boston, MA 02215
  (617) 353-7857,7858
--
(O)((O))(((O)))((((O))))(((((O)))))(((((O)))))((((O))))(((O)))((O))(O)
(O)((O))(((               slehar at park.bu.edu               )))((O))(O)
(O)((O))(((    Steve Lehar Boston University Boston MA     )))((O))(O)
(O)((O))(((    (617) 424-7035 (H)   (617) 353-6741 (W)     )))((O))(O)
(O)((O))(((O)))((((O))))(((((O)))))(((((O)))))((((O))))(((O)))((O))(O)



More information about the Neur-sci mailing list