一种用于块运动估计的匹配准则函数——块特征匹配函数  

A NEW MATCHING CRITERION FUNCTION FOR BLOCK MOTION ESTIMATION-BLOCK FEATURE MATCHING FUNCTION

作　　者：骆立俊[1] 邹采荣[1] 何振亚[1] 

机构地区：[1]东南大学无线电工程系,南京210096

出　　处：《电子科学学刊》1998年第4期486-491,共6页

基　　金：国家攀登计划(神经网络)重大项目;东南大学与菲立浦公司消费电子领域合作项目资助课题

摘　　要：本文提出了一种新的块运动估计匹配准则函数——块特征匹配(BFM)函数,可以用于视频压缩的一些国际标准,如H.261,H.263,MPEG1,MPEG2,HDTV的编解码器中。在这些视频压缩国际标准中视频系统编码器的复杂性最主要取决于运动估计算法。实时的块匹配运动估计的VLSI实现需要考虑以下几个方面:在给定搜索域内运动搜索的复杂度;每次块匹配运算的匹配计算复杂度;每次块匹配运算需要从帧存读取到运动估计处理器的数据量大小;实时硬件实现的适用性.仿真表明BFM算法非常简单有效,可以大大降低相应的块匹配计算复杂度、匹配运算时数据传输时间.BFM函数便于并行实现,从而可以有效地缩短视频编码器的编码时间。本文还详细地给出了BFM函数与其它常用匹配准则函数的比较结果.A block feature matching (BFM) criterion function used for block motion estimation in some international standards such as H.261, H.263, MPEG1, MPEG2, and HDTV codecs is proposed in this paper. The complexity and VLSI implementation of real-time block matching motion estimation need to involve the following considerations: the motion searching complexity within a given search window, the basic matching complexity for each matching operation, and the required amount of data fetching from frame buffer to the motion estimation processor. Simulation indicated that the proposed BFM function is very simple and effective, thus the corresponding complexity of block matching computation and the time of data fetching are greatly reduced. Moreover, the BFM function can provide the potential of parallel VLSI implementation. Comparisons of the BFM with other criterion functions are also given in detail.

关 键 词：运动估计 MPEG1 MPEG2 匹配准则 视频编码技术 

分 类 号：TN941.4[电子电信—信号与信息处理] TN941.5[电子电信—信息与通信工程]