一种新的运动搜索算法

A Novel Search Algorithm for Fast Block Motion Estimation

根据序列图像中运动矢量的中心偏置分布特性和矢量间的时空相关性,设计了一种基于连续消除的正方形--菱形搜索策略。在此基础上,融合了搜索起点预测和自适应搜索等技术,提出了一种可预测搜索起点的自适应的基于连续消除的正方形--菱形搜索算法(PA-SEA-SDS)。实验表明,该算法以较小的代价取得了与全搜索算法相当的效果,在搜索次数和精度上均优于菱形和正方形-菱形搜索算法。

Based on the cross-center-biased motion vector distribution characteristic of the image sequences and high space-time correlation of adjacent blocks＇ motion vectors, Square-Diamond search strategy based on successive elimination is designed. With this and in confusion of the prediction of initial search point and adaptive search modes, a predicable adaptive Square-Diamond search algorithm using successive elimination （PA-SEA-SDS） is proposed. The experiment results show that this method could obtain almost the same PSNR as full search algorithm at very low computation cost. It outperforms both the diamond search algorithm and square - diamond search algorithm at search points and search precision.