A neural network approach for stereo vision

Author

Mousavi, M.S. ; Schalkoff, R.J.

Author_Institution

Dept. of Electr. & Comput. Eng., Clemson Univ., SC, USA

fYear

1990

fDate

1-4 Apr 1990

Firstpage

808

Abstract

A method for achieving stereo vision using a neural network to solve the correspondence problem is presented. The algorithm is edge based and uses the epipolar constraint. The algorithm is in two stages. The first stage is designed to extract the features or primitives for matching, using a static connectionist network. A similarity of measure is defined for each pair of primitive matches, which are then passed on to the second stage of the algorithm. The purpose of the second stage is to turn the difficult correspondence problem into a constraint satisfaction problem by imposing some relational constraints. This is solved using a network of neurons. The results of computer simulations are presented to demonstrate the effectiveness of the approach

Keywords

computerised pattern recognition; computerised picture processing; neural nets; constraint satisfaction problem; correspondence problem; edge-based algorithm; epipolar constraint; feature extraction; neural network; pattern matching; primitive extraction; relational constraints; static connectionist network; stereo vision; Cameras; Computer simulation; Computer vision; Data mining; Feature extraction; Labeling; Layout; Neural networks; Neurons; Stereo vision;

fLanguage

English

Publisher

ieee

Conference_Titel

Southeastcon '90. Proceedings., IEEE

Conference_Location

New Orleans, LA

Type

conf

DOI

10.1109/SECON.1990.117929

Filename

117929