Motion Estimation Algorithm Analysis For H.264

The emerging H.264 advanced video coding (H.264/AVC) standard achieves significant improvement in coding efficiency compared with the previous standards. This standard can obtain excellent coding efficiency and outperforms the MPEG-4 in the low bit rate environment. So it can meet the requirement of the network-video with low bandwith and high quality. However, it also leads to higher computational complexity at the encoder mainly due to the variable-block-size motion estimation, which also leads to higher computational complexity. This is mainly due to the fact that the H.264 encoder employs more complicated approaches and features to efficiently improve the coding performance. At the same time, the encoder complexity is tremendously increased with these new approaches. Hence, it is an essential research topic to reduce the high encoding complexity while maintaining the good coding performance, especially when it comes to real-time video coding applications.Motion estimation is one of the core techniques of video coding. Motion estimation and motion compensation can reduce the large amount of temporal redundancy that exists between frames of video sequences, which leads to high compression. The research on looking for a motion estimation algorithm that can get an effective and accurate motion vector quickly becomes a hot topic at present.By studying the existing rapid motion estimation algorithms, we proposed three algorithms for motion estimation in H.264: A fast predictive variable-blocksize motion estimation algorithm for H.264/AVC. A predictive block-size filtering algorithm to extraordinarily improve the encoding efficiency in H.264. Hierarchical filtering schemes for inter mode decision in H.264.The first algorithm takes advantage of three effective predictive schemes -stationary block prediction, predictive search for non-stationary blocks and predictive multi-pattern refinement search in merging process. The second algorithm mainly takes advantage of two efficient predictive methods: one is predictive skipping checking the sub-macroblock-level modes according to the sum of absolute difference (SAD) of each macroblock and the other is filtering some sub-macroblock-level modes with an adaptive threshold obtained fromthemode information in the previous determinedmacroblock. According to the temporally stationary characteristics inherent in video sequences, The third algorithm uses three hierarchical filtering schemes: macroblock-level modes filtering, sub-macroblock-level modes filtering and the most time-consuming intra4 × 4 mode filtering based on the distribution of the stationary blocks and an adaptive threshold.Experimental results and comparative analysis are given to demonstrate that our proposed algorithms can effectively fiter and exclude some unlikely candidate modes and can achieve a fast speed-up factor compared with the current Fast Full Search algorithm in JM8.5. Moreover, our algorithms can effectively control the PSNR loss and the bit rate increase.