用于视频编码的维纳滤波器抽头系数表示新方法  

A New Description Algorithm for the Tap Coefficients of Wiener Filter in Video Coding

作　　者：李志斌[1] 常义林[1] 杨付正[1] 刘晓仙[1] 

机构地区：[1]西安电子科技大学综合业务网理论与关键技术国家重点实验室,西安710071

出　　处：《西安交通大学学报》2010年第10期77-82,共6页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(60772134;60902081;60902052);高等学校学科创新引智计划资助项目(B08038)

摘　　要：针对已有视频编码系统中自适应环路滤波器抽头系数编码方法存在冗余的问题,提出一种基于直流补偿的抽头系数编码表示方法(DCTCC).首先,对解码恢复图像与原始图像的直流差值进行补偿;然后,以直流补偿后的恢复图像为输入信号、原始图像为期望值设计维纳滤波器,进一步补偿恢复图像的交流失真;最后,仅编码传送所设计滤波器的各非中心抽头系数,而根据提出的抽头系数之和等于1的特点则可计算得到中心抽头系数数值,从而提高滤波器参数传送效率.实验数据表明,对于所用的各种格式视频测试序列,DCTCC的编码效率均优于已有自适应环路滤波方法,与H.264/AVC标准相比,DCTCC的编码效率最高可达26.31%.A novel tap coefficients codec method based on DC offset（DCTCC）is proposed to overcome the redundant shortcoming in describing tap coefficients of adaptive loop filter in the video coding.Firstly,DCTCC performs compensating on the DC offset between the restored picture and the original picture.Then,the Wiener filter is designed in terms of the compensated picture as the filter input and the original picture as the desired response to further compensate AC distortion in restored picture.Finally,only the non-center tap coefficients are coded and transmitted,while the value of the center tap coefficient is derived by the feature that the sum of the tap coefficients equals 1,so that the transmission efficiency of the filter parameters is improved.Experimental results with test sequences in different formats show that DCTCC outperforms the existing algorithm with adaptive loop filter.Comparisons with H.264/AVC show that the coding efficiency of DCTCC has an improvement up to 26.31%.

关 键 词：自适应环路滤波器 维纳滤波器 抽头系数 直流补偿 

分 类 号：TN919.8[电子电信—通信与信息系统]