H.264块匹配运动估计中UMHexagonS算法的优化

Optimization of UMHexagonS on block matching motion estimation algorithms in H.264

块匹配运动估计是视频压缩中的核心技术之一。研究高效的、快速的运动估计算法是目前视频压缩技术中研究的重要课题。通过对视频编码标准H.264采纳的UMHexagonS算法分析的基础上,对UMHexagonS算法中的终止搜索和搜索模板这两个部分进行了优化,有效降低了搜索点数,提高了运动估计算法效率。通过对四个典型的测试序列的实验证明,优化后的算法与原算法相比较,在图像质量几乎没有改变的情况下,运动估计时间平均减少了11.75%,提高了视频压缩实时性。

Block matching motion estimation is one of the core-technique of video coding.Research on efficient and fast video compression coding technology has already been a hot topic for technician and researchers all over the world.Based on the analysis of UMHexagons algorithm adopted by video compression H.264 the coding standard,the termination search and search template have been optimized respectively,which reduces the search points effectively,and increases the efficiency of motion estimation algorithm.The experiments on four typical video sequences prove that,compared to the original algorithm,the motion estimation time can be saved 11.75% on average by the optimized algorithm,and the real-time performance of the video compression has been enhanced under almost the same circumstance of picture quality.