一种改进的三步搜索块运动估计算法  

An Improved Three-Step Search Algorithm for Block Motion Estimation

作　　者：谢琅[1] 杨艳[1] 曹阳[1] 王立恒[1] 

机构地区：[1]武汉大学电子信息学院,湖北武汉430072

出　　处：《武汉大学学报（理学版）》2005年第5期625-628,共4页Journal of Wuhan University:Natural Science Edition

基　　金：国家自然科学基金资助课题(60132030)

摘　　要：针对目前块运动估计的三步搜索法和新三步搜索法在搜索时容易陷入局部最小的问题,提出了一种改进的三步搜索法.该算法在三步搜索法的基础上,设计了新的大小菱形综合搜索模板,采取了自适应选择模板的搜索策略,减小了搜索时陷入局部最小的概率,提高了运动估计精度且降低了计算复杂度.模拟实验结果表明,该算法与原三步搜索法相比,运动估计的均方误差更小(降低了0.7)且计算复杂度减小了10%以上.The three-step search algorithm and the new three-step search algorithm both easily lead the search to be trapped into a local minimum. In order to solve this problem, an improved three-step search algorithm is proposed in this paper. Based on the three-step search algorithm, this algorithm employs the diamond-search＇s large and small diamond pattern. Experiment results show that the improved three-step search performs better than former three-step search in terms of Minimum Square Error （reduces about 0. 7） and requires less computation by up to about 10%.

关 键 词：块匹配算法 运动估计 视频编码 

分 类 号：TP919.81[自动化与计算机技术]