基于运动类型与多模板的运动估计算法  

Motion estimation algorithm based on motion classification and multi-pattern

作　　者：潘琢金[1] 张俊祥[1] 罗振[1] 杨华[1] 

机构地区：[1]沈阳航空航天大学计算机学院,辽宁沈阳110136

出　　处：《计算机工程与设计》2015年第11期3060-3063,3101,共5页Computer Engineering and Design

基　　金：航空科学基金项目(2014ZC54012);辽宁省自然科学基金项目(2013024002);辽宁省教育厅基金项目(L2013063)

摘　　要：在UMHexagonS算法的基础上,提出一种基于运动类型与多模板的运动估计算法,有效解决UMHexagonS算法在运动缓慢的视频序列搜索中存在搜索点数过多、速度慢的问题。结合视频序列的运动特征,根据运动快慢程度对运动块进行划分,针对不同类型的运动块采取不同的搜索模板和策略进行搜索。实验结果表明,在保证图像质量的情况下,该算法可显著缩短运动估计时间,提高编码效率。Based on the algorithm of UMHexagonS,a motion estimation algorithm based on motion classification and multi-pattern was proposed which solved the problem of searching too many points and low searching speed of slow-moving video sequence.Distribution features of the motion vector were used,and motion blocks were categorized into different types according to speed,and then the appropriate pattern and strategies for motion blocks were selected to search.Based on the algorithm,the video transmission on mobile device was realized.The results of experiment show that compared with the UMHexagonS,the proposed algorithm can reduce the time of motion estimation obviously and improve coding efficiency significantly on the premise of ensuring the image quality.

关 键 词：运动估计 运动矢量 准静止块 运动类型 搜索模板 

分 类 号：TN919.81[电子电信—通信与信息系统]