A stereo model based upon mechanisms of human binocular vision

A model for stereo vision, which is based on the human-binocular vision system, is proposed. Data collected from studies of neurophysiology of the human binocular system are discussed. An algorithm for the implementation of this stereo vision model is derived. The algorithm is tested on computer-generated and real scene images. Examples of a computer-generated image and a grey-level image are presented. It is noted that the proposed method is computationally efficient for depth perception, and the results indicate accuracies that are noise tolerant.