A stereo pairs disparity matching algorithm by mean-shift segmentation

A stereo pair reconstruction algorithm is proposed which utilizes color mean-shift segmentation on the reference image and local matching based on windows is employed. The scene structure is modeled by a set of planar surface patches, which are needed to be stored for each segment rather than each pixel. Instead of assigning a disparity value to each pixel, a disparity plane is assigned to each segment by which it can be reduce complexity of computing. Test results show that the presented algorithm can reconstruct the scene depths efficiently.