Depth-based block partitioning for 3D video coding

摘要

3D video is an emerging technology that bundles depth information with texture videos to allow for view synthesis applications at the receiver. Depth discontinuities define object boundaries in both, depth maps and the collocated texture video. Therefore, segmentation of the depth map can be used for more accurate partitioning of the corresponding texture component. In this paper, already coded depth maps are used to increase coding efficiency for texture videos by deriving an arbitrarily shaped partitioning of a texture block based on the collocated depth block. By applying motion compensation to each partition and combining them afterwards by a depth-based segmentation mask, highly accurate prediction signals can be formed that reduce the remaining residual signal significantly. Simulation results show bitrate savings of up to 5.7% for the dependent texture views and up to 2.3% with respect to the total bitrate.