| The H.264/AVC stardard is a new international video coding stardard. To satisfy the needs of flexibility and customizability, the H.264/AVC is consisted of a Video Coding Layer (VCL) and a Network Abstract Layer (NAL). In order to transmit video bit streams efficently over error prone environments as well as to adapt the capability of terminal devices, the Joint Video Team of the ITU-T VCEG and ISO/IEC MPEG standardizes a Scalable Video Coding (SVC) as an extension of H.264/AVC. The modes of scalability are temporal scalability, spatial scalability, and SNR scalability.In SVC, the rate-distortion optimization (RDO) technique, which using an exhaustively search to select the optimal coding mode for each macro block (MB), is used to achieve an optimal trade-off between rate and distortion. This technique achieves the highest possible coding efficiency, but it requires an extremely large encoding time which obstructs it from practical use. It is observed that the mode distribution between the base layer and its enhancement layers has a certain correlation. This paper statistics the modes used in practical video coding and proposes a fast mode decision algorithm for Medium-grain Quality Scalability (MGS), which utilizes not only the mode-distribution correlation between the base layer and enhancement layer but also the correlation between MB and its neighboring MBs. The simulation results show that compared to JSVM9.17 reference software, the proposed algorithm can achieve a time saving up to 60.44% on average with negligible loss of quality and few bit rate increment, which is very helpful for practical use. |
PDF Full Text Request