A Direct Disparity Estimation Technique for Depth Segmentation

This paper concerns acquisition of dense depth maps in the context of image segmentation, which is fundamental in various vision applications. Focusing on stereoscopic vision as the methodology, our goal here is to develop a scheme uihich is computationally s imple but still allouis a dense disparity map. For the estima- tion of stereo disparity, for t his reason, uie employ an approach based on local Fourier phase obtained by com- plex bandpass filters. We consider the characteristics of this search-free and thus fast approach as suitable to -. compute stereo disparity as a basic cue for image seg- m.entation. Within the frameuiork of the phase-based al- gorithm, in this paper? two issues are discmsed. One i s the we of the deriz~ati.ue-based filters. and the other is the certainty-weighted disparity propagation.