Research On Scalable Video Coding Based On DCT

This dissertation makes a comprehensive research on scalable video coding based on DCT coding structure. Mainly including the following content:1．Research on source coding technology of stream mediaIn recent years, the digital video coding technology has been developing continuously. New technologies has been integrated into international standards, and been put into practice quickly. It makes the research on video coding active. With the development of broad-band Internet and 3G, video applications grows rapidly.As far as the end-user is concerned, there are two basic approaches to look on video on networks: by downloading and by streaming. When using downloading, the user has to download the whole video file to local harddisk before using it, which will result in the increasing waiting time with the increasing video file size, and be limited by the bandwidth of networks, while the streaming method makes it possible to download and display at the same time. So the video streaming system became the main approach for video storing and distribution nowadays and for the future.Firstly, the dissertation will introduce the structure of video streaming system, and compare the 3 methods which compose the system. Then describe the transcoding and scalable video coding method in detail.2．Research on scalable video coding based on H.264/AVCWith the development of the international video coding standard H.264/AVC, more and more video applications including stream media have or will adopt it. Because of various networks and devices, the requirements on scalable video coding are increasing rapidly.The scalability could be achieved on both DWT or DCT coding structure. On the one hand, DWT has so-called natural embedded feature, which could achieve the scalability more conveniently. On the other hand, lots of video coding standard based on DCT including H.264/AVC have occupied most of the applications, so SVC based on DCT could be compatible with these standard and be adopted easilier. So the MPEG of ISO and VCEG of ITU-T has agreed to research on SVC together as the extension of H.264/AVC, and the Joint Video Team proposed the Joint Draft of SVC.The scalability requires a functionality to allow the removal of parts of the bitstream while achieving a rate-distortion performance of the video coding at any supported spatial, temporal, or SNR resolution that is comparable to single-layer H.264/AVC coding at that particular resolution. Therefor there are three kinds of scalabilities: temporal, spatial and SNR.The dissertation described the spatial scalability in detail, including the layered structure of spatial scalable coding. Proposed a new fast mode decision method for interlayer prediction of spatial scalable coding, which reduced the computational' complexity significantly while only neglectable loss of RD performance occured.3．H.264/AVC video coding based on DSPAlthough H.264/AVC improves the compression performance significantly, it's cost increases, too, including more computation and more store memory, which has been the main obstacle for H.264/AVC to the applications. When the ASIC of H.264/AVC is not mature enough, multimedia DSPs which are programmable become the optimal choice for embedded device to implement H.264/AVC. The dissertation chose the TMS320DM642 DSP from TI Corporation, using the open-source X264 as the reference software, and implemented the realtime H.264/AVC encoding for CIF sequences.