Multiple Description Video Coding Research

With the development of communication technology, multimediaapplications becoming an integral part in our life, such as video calls, interactivevideo sessions, video conferencing, distance learning and so on. These videocommunications services make people’s life more convenient. However, thelarge amount of video data and complexity codec characteristics make thetraditional packet-switched network communication mechanism cannot meet thereal-time video communications business and strong interaction requirements onwireless communications and low-power mobile multimedia devices. At thesame time, the problem of unstable network bandwidth, network congestion androuting delay is easy to cause video data loss and burst errors. In addition, lowpower mobile multimedia equipment’s life time is short, the processor’sperformance is limited, and it is difficult to meet the traditional video codecalgorithm, it will influence the user’s experience, seriously restricted thedevelopment of video communication. Multiple description video coding(MDC)as a robustness to transmission technique for lossy channel, withoutchanging the existing network infrastructure conditions, improve the robustnessof video transmission, has become a hot academic research at home and abroad.This paper presents three kinds of multiple description video codingschemes, respectively: cross-platform multiple description video coding scheme,multiple description video coding based on distributed compressed sensing andmultiple description video coding based on residual distributed compressedsensing. The main works are listed as follows.First of all, this thesis proposes a cross-platform multiple description videosolution. After investigating the video coding, transport protocols, and GUIdisplay design on different platform, we put forward a multiple description ofvideo coding framework. It has the characteristics of good portability, and thecore module of high rate of code reuse. We conducted performance test andverification in the Linux system. Secondly, the thesis have the low complexity characteristics of distributedencoding into multiple description video coding, presents a multiple descriptionvideo encoding based on distributed compressed sensing. It reduces thecomplexity of video coding and improves the robustness of video transmission.Finally, this thesis takes into account of multiple description videoencoding of recent years, provides multiple description video coding schemebased on residuals distributed compressed sensing. It effectively utilizes theresidual coding, LDPC coding and side information coding technology, tofurther improve the video compression ratio and deduce the complexity of videocoding.