Fast mode decision algorithm in H.264/AVC base on statistical feature
Donghui Zhou, Chunjiang Duanmu · 2010
H.264/AVC uses the Rate Distortion Optimization (RDO) technique to select intra and inter prediction modes in order to achieve coding efficiency, but the huge computational complexity makes it unpractical. In this paper, we propose a statistical feature approach to reduce the computation involved in the fast mode decision algorithm. From statistical analysis, the distribution of modes is not uniform; we can ignore some modes that are almost not used because they have a very little effect on encoding efficiency. Then, some statistical features are presented in advance to estimate the best coding mode early. The proposed algorithm is used to determine the best coding mode according to the presented statistical features. Experiment results show that the proposed algorithm can save more encoding time than the High Complexity Mode decision and Fast High Complexity Mode decision algorithms with negligible degradation of RD performance.