Optimal bit allocation for multiview video plus depth coding based on view rendering distortion model

摘要

The depth map provide per-pixel with depth corresponding to the texture video that can be used to render arbitrary virtual views by using depth map based rendering (DIBR). Consequently, for the efficient transmission of 3D video signals, the compression of texture video and the corresponding depth map is required. Since the bit rate ratio between the texture video and depth map is still an open question, this paper introduces a novel bit allocation algorithm for texture video and depth map based on view rendering distortion model. Firstly, we derive relationship between distortion in coded texture/depth and the rendered view. Then, a novel distortion model is proposed to precisely estimate distortion in view synthesis. Based on this model, a bit allocation algorithm is proposed to optimize the bitrates between texture video and depth map for minimizing the view rendering distortion. Experimental results show the efficiency of the proposed distortion model, and that the proposed bit allocation algorithm can achieve the almost same rate-distortion (RD) performance as the full search algorithm while can significantly reduce the computational complexity. (C) 2015 Elsevier GmbH. All rights reserved.