Multiview Multiple Description Video Coding

With the development of information science and 3D video technology, many studies paid attention to compression efficiency and rate distortion performance of 3D Video. When 3D videos are transmitted over error-prone channels, they may suffer significant quality degradation. In this thesis, we combine multiview video coding (MVC) with multiple description coding (MDC) for robust transmission. The proposed schemes are focus on multiple description coding and multiview coding. Specially, the main contributions of this thesis are as follows:(1) We attempt to introduce multiple description coding (MDC) to 3D video compression for robust transmission. By taking full advantage of both intra-view and inter-view correlation, the proposed scheme makes an adaptive mode decision to generate the labels as redundant information. The redundant information occupies just a few bits, while the PSNR of video sequences reconstructed by the proposed scheme are improved a lot. Therefore, it may be a worthy choice for 3D video applications over error-prone channels.(2) We exploit the characteristics of multiview video by using motion vector, disparity vector and residual information as redundancy. Six encoding models are constructed. Based on the idea of rate distortion optimization, the redundant information is reasonably selected, and the mode selection is carried out. Experimental results show that the proposed scheme can improve the robustness of video transmission. In particular, when errors occur in the transmission process, the use of the proposed can improve the quality of the reconstruction of the decoder. At the same time, redundant information is controlled within a reasonable number, which shows a better rate distortion performance.(3) By using the optimized DCT domain zero padding method, the quality of side reconstruction is improved. And the residual compensation is performed on the video, the quality of the reconstructed image is further improved. Only use part of the original video data, and then make the residual compensation, we can get better quality of the reconstructed video.