基于AVS的UMHexagonS算法研究与改进

Optimization on UMHexagonS Algorithm Based on AVS

运动估计是视频压缩中最重要的环节。该文在分析UMHexagonS算法的基础上，进行了两方面的改进。首先，针对它在高效的起始点预测后陷入搜索冗余的可能，增加了一个提前终止判断。另外，对算法中的大六边形搜索模板进行了改进以减少搜索的点数，在保持原有图象质量的情况下有效的节省了运动估计时间。通过对各种测试序列的实验证明，在保证PSNR和码率变化不大的前提下．可以有效地降低4％-25％的运动估计时间。

Motion estimation （ME） is vital to video compression. In this paper ,based on Analyzing the UMHexagonS algorithm and give two improvement measures.firstly, an early term ination is added in the case that it enters the redundant search after the initial search point prediction with good effects.secondly, ,an optimized algorithm is proposed based on big hexagon,which saved the search point and saves ME time effectively without quality loss. Experimenting with some typical video sequences proves that ,this ME time can effectively reduce 4%-25% while the PSNR and bit Rate in ensuring little chang.