A simple feedforward neural network architecture for 3-D motion and structure estimation

Author

Sun, Yi ; Bayoumi, Mohamed M.

Author_Institution

Dept. of Electr. & Comput. Eng., Queen´´s Univ., Kingston, Ont., Canada

Volume

3

fYear

1996

fDate

16-19 Sep 1996

Firstpage

783

Abstract

A simple feedforward neural network system which is especially designed to tackle the problem of 3-D motion and structure parameter estimation from 2-D optical flow parameters has been proposed. Each node and the weight of every connection adopted in the network has its explicit physical meaning. The network embraces a self tuning scheme with an unsupervised learning rule to control the dynamics of the system. It also adopts a mechanism for preattentative focus that effectively suppresses the spurious solution of the estimation

Keywords

feedforward neural nets; image sequences; motion estimation; neural net architecture; parameter estimation; unsupervised learning; 2D optical flow parameters; 3D motion estimation; 3D structure parameter estimation; feedforward neural network architecture; preattentative focus; self tuning scheme; system dynamics control; unsupervised learning rule; Control systems; Feedforward neural networks; Image motion analysis; Neural networks; Optical computing; Optical design; Optical fiber networks; Optical tuning; Parameter estimation; Unsupervised learning;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 1996. Proceedings., International Conference on

Conference_Location

Lausanne

Print_ISBN

0-7803-3259-8

Type

conf

DOI

10.1109/ICIP.1996.560848

Filename

560848