BMV: Behavioral Model of Active Visual Perception and 
Invariant Recognition
=============================================================
Authors:  Ilya A. Rybak; Valentina I. Gusakova, 
          Alexander V. Golovan, and Natalia A. Shevtsova

BMV was developed in the Laboratory for Neural Network Modeling in 
Vision Research, A. B. Kogan Research Institute for Neurocybernetics 
at Rostov State University (Rostov-on-Don, Russia) in 1991.

This DEMO demonstrates the ability of the model to recognize complex 
gray-level images (e.g. faces) invariantly with respect to 2D 
transformations (shift, rotation, scale).

DEMO was written in DOS and does not require WINDOWS.

To see DEMO you have to put all files (bmv.exe, i1.dat - i6.dat) 
into the same directory and run bmv.exe file.

DEMO consists of three modes (regimes).

I. In the mode of "Image Memorizing", the test images appear on the 
screen and are memorized.  Each image is memorized by sequential 
picking points of interest by shifting an "attention window". 
The individual scanpath of viewing for the processed image is formed 
during this regime (Note that at each fixation the part of image with 
the center at the fixation point is processed with resolution decreased 
with the distance from the fixation point). 

PRESS ANY KEY TO CHANGE THE IMAGE

2. In the next mode you can SELECT AND TRANSFORM the image 
which you wish to present for recognition. 

TO SELECT A REQUIRED OPTION USE ARROWS AND PRESS "Enter"

To select or change the recognized image use "CHANGE".
Use "SCALE", "ROTATE", "TRANSLATE", and "NOISE" in order 
to apply the corresponding image transformations and/or noise.

ATTENTION!   PLEASE, APPLAY THE "SCALE" TRANSFORMATION FIRST.

IN "TRANSLATE" MODE USE ARROWS TO SHIFT THE IMAGE and then
use "Escape" TO EXIT THE "TRANSLATE" MODE.

3. Recognition mode

USE "RECOGNIZE" OPTION TO PRESENT THE SELECTED, TRANSFORMED IMAGE 
TO THE MODEL FOR RECOGNITION. 
The result of recognition appears at the top.

QUESTIONS, SUGGESTIONS, COMMENTS? 
Contact Ilya Rybak 
(rybaki@eplrx7.es.duPont.com)
or 
Alexander Golovan, Valentina Gusakova, Natalia Shevtsova 
(w516@krinc.rnd.runnet.com)

ENJOY!

25/3/1996