Stereo Correspondence Using Hierarchical Belief Propagation and Cross-Based Region

In this paper, we present a new global stereo matching algorithm based on belief propagation and cross-based aggregation region, assuming that neighboring pixels with similar colors share same disparities. In this approach first the hierarchical belief propagation strategy followed by a left-right consis-tency checking is used for initial match. Then a refinement algorithm is applied to disparity map, based on cross-based region. Initial matching performed by hierarchical belief propagation strategy uses less memory at run-time and im-proves the running speed for minimizing the energy function. The experimental results, evaluated by the Middlebury data sets, show that the proposed method improves the accuracy of disparity map compare to other real-time stereo matching algorithms.