A wavelet-based multiresolution approach to stereo matching of face images

Stereo matching is one of the most active research areas in computer vision. To find the corresponding point in a stereo pair of face images often consider intensity or color differences as a local matching metric, which is sensitive to contrast changes. We proposed a multiwavelet based algorithm for stereo matching, which depends on multiple spatial frequency channels for matching of a stereo pair of face image. The stereo test images of popular face database and middlebury database are used to generate experimental results. The results show that the proposed technique produces disparity maps with fewer mismatch errors.