Research On Scalable Video Coding Based On Error-prone Network

Real-time video transmission over IP based network, such as Internet and wireless communication network etc., is developed broadly and deeply in the last years. Video compression coding has been developed one of most important topics in this area. However. The traditional IP based network which is called error-prone only provides best-effort service at network layer from the point of technology generally. The Best-effort mechanism cannot provide efficient guarantee for the network transmission parameters which are essential to those real-time implement such as video streaming over Internet real-timely. It imposes some bad impact on video implement over network and imposes many challenges to the traditional nonscalable video coding technology which is not efficient for transmission video over the error-prone. Then the scalable video coding technology was proposed, which can provide a rang of bitrate change and its redundant video content compensate the transmission error, but loss coding efficiency. Recently the research on video scalable coding based on error-prone network has been one of hot and difficult topics in video compression coding filed today. Based on the analysis of some of current video coding technologies and the error-prone characteristics, the main works and conclusions in this thesis are listed as follows:(1) Hybrid video scalable coding structureBased on the analysis and research of the current several scalable coding schemes, a multi-loop independent motion compensation encoding structure integrated several existing standard scalability is proposed to provide a broader bitrate range to be adaptive the status of the network and to meet requirement of different level of quality of service, and improve the coding efficiency.(2) Interframe prediction modelResearching on the existing video scalable coding structure and the Predict-tion model shows that the information between frames of video sequence was not utilized fully to improve coding efficiency in these coding scheme, especially the prediction of B frame in the enhancement layer. A scheme to improve the B frame prediction based on the proposed scheme was proposed in this thesis. (3) Experiment and resultsAt the end of thesis, the simulation experiments are implemented. The experimental results demonstrate that the proposed video scalable coding scheme can provide a broader bit rate range with a improvement of restructured image quality.