Error Resilience Study Of Scalable Video Coding Based On Wyner-Ziv Framework

With the continuous development of technologies, such as network, all kinds of computations, the demand of video services has become complexity. In this case, the scalable video coding generated to improve efficiency of video coding and service quality. For actual demand of video application, in this thesis we focus on the robustness of the scalable video coding with the video coding mechanism proposed by Wyner-Ziv.The major contributions of this thesis are as follows:1. For Heterogeneous Network Environments, we study the problem of video coding and its transmission, narrate the conceptions of scalable video coding and summarize the principal theory basis and coding methods using different technologys. On the basis of above work, we analyze the advantages and shortcomings of scalable video coding structures. Considering the manifold scalability of wavelet transform coding, we choose scalable video coding structure MC-EZBC as the basis of our error resilience structure.2. From Wyner-Ziv video coding framework, we elaborate the source and theory basis of it. On the foundation of the existing structures of Wyner-Ziv video coding, we analyze the applicability of Wyner-Ziv video coding and the problems of used in scalable video coding.This is useful for our designing error resilience structure.3. The thesis presents a novel error resilience scheme based wavelet scalable video coding structure MC-EZBC. In this encoder system, while video signal be encoded by MC-EZBC structure, we use Wyner-Ziv coding architecture to do error resilience. At the decoder, we consider a coarse estimated version of the corrupted parts from MC-EZBC decoder as the side information to generate a refine version by the Wyner-Ziv description from the auxiliary channel. Comparison with forward error correction, simulation results show that Wyner-Ziv description can effectively correct errors and improve the quality of reconstruction frame.