Pre-/post-processing to improve the coding performance of multiview plus depth map

This paper presents pre- and post-processing schemes to improve the coding performance of a multiview plus depth map (MVD) system, which is one of the key frameworks for next-generation 3D video systems. In previous work, depth maps have been considered as a different type of video, and a conventional coding scheme is simply applied. The scheme proposed here instead focuses on the unique properties of depth maps and utilizes them to improve the coding performance of MVD systems. In detail, the pre-processing stage selectively filters the input depth map to reduce the bit-rate cost while minimizing quality degradation, and the post-processing restores the reconstructed depth map as much as possible. Additionally, we propose an optimal filter parameter selection scheme obtained using a theoretic and experimental analysis of the rate-distortion relationship. The experiments show that the proposed approach improves the performance of an MVD system.