一种改进的运动矢量编码方法  

An advanced algorithm for motion vector coding

作　　者：哈力旦.A 方勇[2] 

机构地区：[1]新疆大学电气工程学院,新疆维吾尔自治区乌鲁木齐830008 [2]西安电子科技大学综合业务网国家重点实验室,陕西西安710071

出　　处：《西安电子科技大学学报》2005年第4期639-642,共4页Journal of Xidian University

摘　　要：对现有的运动矢量编码算法作了改进.新算法根据图像在水平及垂直方向运动量的差异,将运动矢量的编码分为3种模式.对不同的编码模式采用不同的码表进行编码.仿真结果表明,新算法减少了10%～40%的运动矢量编码比特数,同时提高了重建图像质量达0.05～0.1 dB.此外,新算法还提高了输出码流的抗误码性能和编码器流控的精确程度.Some modifications are made on the present algorithm for motion vector coding. In the new algorithm,motion vector coding is divided into three modes according to the difference between horizontal movement and vertical movertmnt. For different coding modes, different variable length code （VLG） tables are used. The simulation results show that the new algorithm reduces 10% ~ 40% bits for motion vector coding and improves the quality of reconstructed image by 0.05- 0. 1 dB. In addition, the new method improves the robusness of video bit streamagainst channel errors and makes the rate controlling operation of video encoders more accurate.

关 键 词：MPEG-4 PFGS 流媒体 网络视频 运动估计补偿 

分 类 号：TN919.81[电子电信—通信与信息系统]