基于运动估计的置信度传播立体视频匹配算法

Stereo Video Matching Algorithm of Belief Propagation Based on Motion Estimation

基于置信度传播(BP:Belief Propagation)的立体匹配算法与局部算法相比,视差准确度高,但需要耗费大量的计算时间。为此,提出一个基于运动估计的置信度传播立体视频匹配算法。该算法首先通过传统的BP算法获得I帧的匹配视差图,并保存每个像素获得最佳视差值的传递信息;然后,通过参考I帧的运动估计信息,得到重新排列的I帧视差值的传递信息,将其作为P帧置信度传播算法的初值进行迭代运算,从而大大减少了P帧置信度传播算法的迭代次数。实验结果表明,该算法能大大提高置信度传播立体视频的匹配效率。

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.