摘要
研究运动估计(ME)中的经典块匹配算法(BMA)并重点分析菱形(DS)算法,然后利用相邻块之间空间关系,设计一种新的块匹配策略,以相邻宏块的搜索结果进行预测匹配,以减少搜索点数。最后以平均每块搜索点数和平均峰值信噪比为衡量指标验证了该算法的有效性。
The paper studies the classical fast block-matching algorithm （BMA）in motion estimation （ME）and pays attention to analy-zing diamond search （DS）algorithm.Next by taking advantage of the neighbor contact between neighboring blocks,it designs a new block matching strategy which takes the search result of neighboring macro blocks to predict the match so that the number of search points can be reduced.Finally by taking both average search points of each block and average PSNR as measurements,the algorithm is validated.
出处
《计算机应用与软件》
CSCD
北大核心
2014年第2期164-166,170,共4页
Computer Applications and Software
关键词
块匹配
运动估计
菱形算法
视频压缩
Block match
Motion estimation
Diamond algorithm
Video compression
