Fast Inter Mode Decision Algorithm for H.264/AVC Based on Motion Complexity Analysis

摘要

In H.264/AVC video standard, inter-mode prediction technique, is used for reducing temporal as well as spatial redundancies by employing seven variable size of MacroBlocks (MBs). To determine the best MB among the different sizes the rate distortion cost is used as a determining factor. This process requires relatively large amount of computation to encounter all the combinations of different sizes of Mbs. Hence, a fast inter-mode decision algorithm is required to reduce the computational complexity in H.264/AVC. In this study, a new fast inter-mode decision algorithm is proposed in this context. The proposed method reduces the number of candidate modes using direct information from the co-located macroblock (s). A novel motion complexity analysis scheme is also discussed in this study which is strong enough to categorize different moving objects inside a MB and to make a proper decision for the best candidate mode. Experiment results show that the proposed algorithm can achieve speed-up factor up to 78.70% on average with negligible loss of image quality and quite good bit rate performance.