Joint Bit Allocation and Rate Control for Coding Multi-View Video Plus Depth Based 3D Video
Feng Shao,Gangyi Jiang,Weisi Lin,Mei Yu,Qionghai Dai
DOI: https://doi.org/10.1109/TMM.2013.2269897
IF: 7.3
2013-01-01
IEEE Transactions on Multimedia
Abstract:In three-dimensional (3D) video coding, distortion in texture video and depth maps can all affect the quality of the synthesized virtual views. Therefore, under the total bitrate constraint, effective bit allocation between texture and depth information is very important for 3D video coding. In this paper, the major technical contribution is to formulate view synthesis quality for optimal resource allocation in 3D video coding, since such quality is what that matters most to the ultimate user (i.e., the viewer) of the system; to be more specific, a new joint bit allocation and rate control method for multi-view video plus depth (MVD) based 3D video coding is proposed accordingly. We firstly derive a view synthesis distortion model to characterize the effect of coding distortion of texture video and depth maps on the synthesized virtual views. Based on this model, we derive a rate-distortion model to characterize the relationship between the bitrate and the view synthesis distortion, and the optimal bitrate ratio between texture and depth is established adaptively by solving the associated optimization problem. Finally, the rate control algorithm is performed on view level, texture/depth level and frame level. Experimental results show that compared with other methods, the proposed bit allocation method obtains higher performance of view synthesis. Moreover, the proposed rate control method can accurately control the bitrate to satisfy the total bitrate constraint.