Method and apparatus for determining image affine flow using artifical neural system with simple cells and lie germs
First Claim
1. An artificial neural system for extracting local image affine flow information from visual stimuli comprising:
- a) time differentiation means for receiving time-varying image intensity signal and providing image time-derivative signal;
b) hypercolumn (HC) reference frame means coupled to said time differentiation means for receiving said image time-derivative signal and providing HC-time-derivative signal;
c) hypercolumn (HC) Lie differentiation means for receiving image intensity signal and providing six HC-Lie-derivative signals;
d) least square error fitting means coupled to said HC-reference frame means and said HC-Lie differentiation means for receiving HC-time-derivative signal and six HC-Lie-derivative signals and determining local affine flow.
0 Assignments
0 Petitions
Accused Products
Abstract
A method and apparatus is provided to determine image affine flow from time-varying imagery. The novel artificial neural computational system of a cortical hypercolumn comprising a plurality of specific orientation (SO) columns and a least square error fitting circuit is based on a Lie group model of cortical visual motion processing. Time-varying imagery, comprising intensity imagery and time-derivative imagery is provided to a plurality of specific orientation (SO) columns comprising simple cells and Lie germs. The cortical representation of image time derivative and affine Lie derivatives are extracted from responses of simple cells and Lie germs, respectively. The temporal derivative and affine Lie-derivative information obtained from each specific orientation (SO) columns is applied to least square error fitting analog circuit having a three layer multiplicative neural architecture to determine image affine flow components in accordance with an error minimization gradient dynamical system technique.
-
Citations
20 Claims
-
1. An artificial neural system for extracting local image affine flow information from visual stimuli comprising:
-
a) time differentiation means for receiving time-varying image intensity signal and providing image time-derivative signal; b) hypercolumn (HC) reference frame means coupled to said time differentiation means for receiving said image time-derivative signal and providing HC-time-derivative signal; c) hypercolumn (HC) Lie differentiation means for receiving image intensity signal and providing six HC-Lie-derivative signals; d) least square error fitting means coupled to said HC-reference frame means and said HC-Lie differentiation means for receiving HC-time-derivative signal and six HC-Lie-derivative signals and determining local affine flow. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. In an artificial neural system including time differentiator means for receiving time-varying intensity image signal and providing time-derivative image signal, HC-reference frame means for receiving time-derivative image signal and providing HC-time-derivative signal, HC-Lie differentiation means for receiving intensity image signal and providing HC-Lie derivative signals, least squares error fitting means for receiving said HC-time-derivative signal and HC-Lie-derivative signals and providing affine flow signal, the method for determining the affine flow of the time-varying image comprising the steps of:
-
differentiating said time-varying intensity image by time-differentiation means; extracting HC-time-derivative Ω
t from said time-derivative image by using SO-simple cells in said HC-reference frame to obtain each said SO-simple cell response signals;extracting HC-Lie-derivatives Ω
1, Ω
2, Ω
3, Ω
4, Ω
5, Ω
6 SO-Lie-germ response signals from said intensity image by using said SO-Lie germs of said Lie derivative operators to obtain each said SO-Lie-germ response signals;applying said HC-time-derivative and said six HC-Lie-derivatives to respective least squares error fitting circuit to explicitly determine six affine flow component parameters ρ
1, ρ
2, ρ
3, ρ
4, ρ
5, and ρ
6. - View Dependent Claims (17, 18, 19, 20)
-
Specification