Research On The Fast Prediction Algorithm Of Scalable Video Coding

Along with the rapid development of wireless communication, the application of the multimedia services becoming more widely in the field of mobile communication, especially streaming media services as the main content of video, has been widely used in many fields such as video conference, mobile television, visual phone call. To adapt the variable condition of network transmission and the diverse requirements of different clients, the international organisation for standardisation ISO/IEC MPEG and ITU-T VCEG form the Joint Video Team(JVT) and develop a Scalable Video Coding(SVC) standard as the amendment for the H.264/AVC standard. SVC system can achieve tamporal, spatial and quality scalability with one time coding process. However, as the using of new coding tools and layered structure, it introduce high coding complexity and computation, which obstruct practical application of SVC.To solve problem of high encoding complexity in SVC, we propose two fast prediction alogrthms with diffenent scalable structure. First, in terms of intra prediction, a fast inter-Layer intra prediction algorithm is proposed by making use of the mode correlation between coding layers to improve the encoding efficient. By analyze the mode distribution of encoding layers, this algorithm encode the enhancement-layer base on the mode information of base-layer to reduce the number of candidate mode. Second, in terms of inter-frame prediction, a fast mode decision algorithm for inter-frame prediction has been proposed in tamporal scalability by make used of correlationship between different tamporal layers. Experimental results show that both of the proposed algorithms can save more than 30% encoding time on average with negligible change of PSNR and bit-rate, this improve the coding efficiency.