多分辨率运动估值的改进算法  

An Improved Multiresolution Motion Estimation Algorithm

作　　者：刘红梅[1] 黄金科[1] 肖自美[1] 梁凡[1] 

机构地区：[1]中山大学电子与通信工程系,广东广州510275

出　　处：《中山大学学报（自然科学版）》2001年第2期34-38,共5页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金资助项目!( 699750 1 1 );广东省自然科学基金资助项目!( 980 4 42 );广东省自然科学基金重点资助项目!( 963 0 3

摘　　要：多分辨率运动估值和补偿算法是一种广泛应用于小波视频编码器的消除时间冗余的算法 ,针对其运动矢量编码效率低及采用全搜索方法、运算复杂度较高等问题 ,提出了一种基于区域分割的多分辨率运动估值算法 .该算法具有如下特点 :①充分利用了小波分解后各子带所表征的运动结构的相关性 ,有效地提高了运动矢量的编码效率 ;②通过对视频帧小波分解后的低通子图像进行静止和运动区域分割 ,以分割结果指导运动估值过程 ,在有效地减少运动估值时间消耗的同时 ,进一步提高了运动矢量的编码效率 ;③在码率受限的情况下 ,重建图像的质量得到提高 .实验结果表明文章所提出算法是有效的 .Multiresolution motion estimation and compensation is a widely used algorithm to reduce the temporal redundency in the wavelet video encoder. Aimed at resolving the problem of the motion vector coding efficiency and the computation complexity,this paper presents a region_segmentation_based multiresolution motion estimation(MRME) algorithm. The features of the presented algorithm are as follow: ① by fully using the spacial and orientation correlations between the motion structure represented by the subbands of wavelet decomposition, the coding efficiency of motion vector is improved;② by segmenting the lowest frequency subimage of wavelet decomposition of the frame into static and motive regions,the computation complexity is reduced effectively and the motion vector coding efficiency is improved farther;③ when the data rate is limited, the performance of codec using the presented algorithm is improved. The experimental results show the effectiveness of the presented algorithm.

关 键 词：小波分解 视频编码 多分辨率运动估值 运动检测 图像压缩编码 编码效率 

分 类 号：TN919.81[电子电信—通信与信息系统]