Interview rate distortion analysis-based coarse to fine bit allocation algorithm for 3-D video coding

In three dimensional video coding (3-DVC), it is reasonable to allocate bits for texture videos and depth maps differently as the characteristic of texture videos is different with that of depth maps. In order to improve the accuracy of bit allocation performance for 3-DVC, a more accurate distortion model for virtual view and rate models for texture videos and depth maps are proposed based on interview rate distortion analysis. Based on the proposed models, the bit allocation optimization problem is resolved by a coarse to fine strategy. The proposed bit allocation algorithm is implemented in the High Efficiency Video Coding-based 3-DVC platform version 6.0 (3-D-HTM 6.0). Experimental results demonstrate that the proposed distortion and rate models are accurate; meanwhile, the absolute difference (denoted as "rate inaccuracy") between the actual coding bit rate (RABR ) and the target coding bit rate (RTBR ) of the proposed method is only 2.67% on average; while the rate inaccuracy of the existing planar model-based method is 17.71%. Besides, there is an average 29.68% complexity increment when comparing with the planar model-based method.