Dense motion and disparity estimation via loopy belief propagation

We describe a method for computing a dense estimate of motion and disparity, given a stereo video sequence con-taining moving non-rigid objects. In contrast to previous approaches, motion and disparity are estimated simultane-ously from a single coherent probabilistic model that cor-rectly accounts for all occlusions, depth discontinuities, and motion discontinuities. The model is a Markov random field (MRF) whose label space incorporates every possible oc-clusion status for every pixel. Hence, the MRF’s data likeli-hoods are physically realistic. The results demonstrate that simultaneous estimation of motion and disparity is superior to estimating either in isolation, and show the promise of the technique for accurate, probabilistically justified, scene analysis. 1. Motivation an