一种改进的H.264视频防失真漂移隐写算法  

An Improved H.264 Video Steganographic Algorithm Against the Drift Distortion

作　　者：王丽娜[1,2] 叶猛[1] 翟黎明[1] 邵强[1] 任延珍[1,2] 王旻杰[1] 

机构地区：[1]武汉大学空天信息安全与可信计算教育部重点实验室,湖北武汉430072 [2]武汉大学计算机学院,湖北武汉430072

出　　处：《武汉大学学报（理学版）》2015年第1期34-40,共7页Journal of Wuhan University:Natural Science Edition

基　　金：教育部博士点基金优先发展领域项目(0110141130006);国家自然科学基金(61103219)资助项目

摘　　要：数字视频中量化后的残差DCT(离散余弦变换)系数常用来作为隐写嵌入的位置,而直接在DCT域上隐写会导致视频帧出现失真漂移现象,使视频质量受到严重破坏.本文在已有文献的基础上,针对H.264/AVC视频,研究了基于纹理特征的帧内8×8亮度块上的隐写模型,将8×8块和4×4块上的隐写模型相结合,提出了一种改进的H.264视频防失真漂移隐写算法.与已有文献中的隐写算法相比,本文提出的隐写算法在PSNR平均仅降低1.94dB的情况下,隐藏容量平均提升30%.The quantized residual DCT coefficients of digital video are often used as steganographic embedding locations, however, embedding directly in DCT domain may cause distortion drift in video frames, which leads to severely damaged video quality. In this paper, based on the existing literature, aiming at H. 264 / AVC video, the texture features based 8×8 luminance block steganographic model is researched, and an improved anti-distortion drift steganography algorithm which combined with the original 4 ×4 blocks steganographic model is proposed. Compared with the existing algorithm, the proposed algorithm can increase the embedding capacity by an average of 30% while the average PSNR loss is only 1.94 dB.

关 键 词：H.264/AVC 失真漂移 帧内预测 耦合系数对 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]