论文部分内容阅读
本文提出了一种块特征匹配预测搜索ＢＦＭＰＳ算法，可以用于视频压缩的一些国际标准，如Ｈ．２６１，Ｈ．２６３，ＭＰＥＧ１，ＭＰＥＧ２，ＨＤＴＶ中．ＢＦＭＰＳ算法充分利用了序列图像的实际运动矢量与预测矢量之间距离的空间分布特征中心偏置分布特性和时间上的相关特性，并采用了中止判决准则，可以明显地减少运动搜索复杂度．ＢＦＭＰＳ算法在匹配运算中采用了简单有效的块特征匹配准则函数，相应的块匹配计算复杂度、数据读取复杂度和内存管理复杂度大大降低．仿真表明这种算法减少了搜索次数，提高了搜索效率，降低了运动估计总的计算复杂性．本文还详细地给出了ＰＳＡ算法与其它常用快速搜索算法的比较结果．
In this paper, a block feature matching predictive search (BFMPS) algorithm is proposed, which can be used for some international standards of video compression, such as H. 261, H. 263, MPEG1, MPEG2, HDTV. The BFMPS algorithm makes full use of the spatial distribution characteristics of the distance between the actual motion vector and the prediction vector of the sequence image, the central bias distribution and the temporal correlation characteristics, and adopts the termination criterion to reduce the complexity of motion search obviously. The BFMPS algorithm uses a simple and effective block matching rule function in the matching operation. The corresponding block matching calculation complexity, data reading complexity and memory management complexity are greatly reduced. Simulation shows that this algorithm reduces the search times, improves the search efficiency and reduces the total computational complexity of motion estimation. The paper also gives a detailed comparison between PSA algorithm and other commonly used fast search algorithms.