Robust approach for disparity estimation in stereo vision

摘要

In this paper, we present a robust probabilistic method for the estimation of stereo disparity. It is based in Bayesian estimation theory, with a prior Markov random field model for the assigned disparities. The optimal estimator is computed using a Gauss-Markov random field model for the corresponding posterior marginals, which results in a diffusion process in probability space. This process, with the appropriate boundary conditions, is also used to estimate disparity in problematic regions of stereo pairs, such as occluded areas and non-textured (homogeneous) regions. Experimental comparisons of the proposed approach with other state-of-the-art methods are presented as well.