An efficient hierarchical disparity estimation algorithm for stereoscopic video coding

Disparity estimation is an important part of stereoscopic video coding. An efficient disparity estimation algorithm is proposed which is based on two successive steps. The former is performed by multiresolution block matching algorithm. The latter consists of a region merging process, which can merge different size blocks with similar disparity vectors into a larger region. The results indicate that the algorithm can overcome conventional algorithm's shortcoming and reduce the coding overhead for disparity information The method is particular suited for low-bit-rate stereoscopic video coding.