适用于MPEG-4的高效空域自适应图像后处理算法  

An efficient spatial adaptive image post-processing algorithm for MPEG-4 video

作　　者：喻庆东[1] 周莉[1] 朱伟成[1] 陈杰[1] 

机构地区：[1]中国科学院微电子研究所,北京100029

出　　处：《哈尔滨工程大学学报》2012年第11期1440-1446,共7页Journal of Harbin Engineering University

基　　金：国家863计划资助项目(2009AA011700)

摘　　要：为有效抑制重建视频图像中的人为编码噪声,提出了一种适用于MPEG-4的具有5种滤波模式的高效空域自适应图像后处理算法.通过对各种人为编码噪声的特性分析,根据待滤波区域像素的一维变化特性和像素的变化范围,将待滤波边界划分为不同的区域,并结合人类视觉系统(HVS)的视觉特性,对不同区域分别采取各种有效的局部特征自适应的滤波策略.算法采用了与其他视频编码标准中环路滤波算法相似的算法结构,便于进行兼容多标准的视频解码器设计.实验结果表明,本算法可有效抑制多种人为编码噪声,对于各种不同特性和不同比特率的测试序列均有良好的可靠性和稳定性,相比于传统算法,本算法可获得更高的峰值信噪比增益.In an attempt to effectively suppress the annoying coding artifacts in reconstructed video frames, an effi- cient spatial adaptive image post-processing algorithm for MPEG-4 video is proposed. By analyzing the characteris- tics of various coding artifacts, the boundaries to be filtered are categorized into different regions based on the 1-D variation properties and variation range of the pixels in the area to be filtered. Combined with the properties of hu- man vision system （ HVS）, various adaptive filtering strategies for local characteristics are effectively applied to dif- ferent regions. In order to facilitate the video codec design for multiple standards, an algorithm structure similar to the in-loop de-blocking filters in other video coding criteria is adopted. Experimental results indicate the proposed algorithm effectively suppresses varied coding artifacts. As well, the results show excellent stability and reliability on the varying test sequences that have different image features and bit rates, obtaining better peak signal noise ratio （PSNR） gain and visual quality, as compared with other traditional algorithms.

关 键 词：图像后处理 人为编码噪声 H 264 AVC环内滤波 图像边缘复杂度判定 自适应线性滤波器 

分 类 号：TN919.8[电子电信—通信与信息系统]