Abstract:

Stereo matching algorithms based on  BP(Belief Propagation) can yield better matching accuracy over the local methods, but large quantity of calculating is needed. For this problem, this paper presents a stereo video matching algorithm based on motion estimation. The traditional BP algorithm is used to get the disparity map of I frame, and the propagating messages with optimal disparity of each pixel are stored. Then, for P frame, the rearranged propagating messages through referring to motion estimation information from I frame are used as the initial values for iteration of BP algorithm. This reduced the number of iterative times. Experiment results show that the proposed stereo video matching algorithm of belief propagation based on motion estimation can dramatically enhance the efficiency of stereo video matching.

Key words: stereo video matching, motion estimation, belief propagation(BP);, markov random field(MRK)