一种基于H.264/AVC的视频可逆脆弱水印算法  

A Reversible Fragile Video Watermarking for H.264/AVC

作　　者：张维纬[1,2] 张茹[1,2] 刘建毅[1,2] 伍淳华[1,2] 钮心忻[1,2] 杨义先[1,2] 

机构地区：[1]北京邮电大学信息安全中心,北京100876 [2]北京邮电大学灾备技术国家工程实验室,北京100876

出　　处：《电子与信息学报》2013年第1期106-112,共7页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(61003284);北京市自然科学基金(4122053);中央高校基本科研业务费专项资金(BUPT2011RC0210);新闻出版重大科技工程(GXTC-CZ-1015004/09;GXTC-CZ-1015004/15-1)资助课题

摘　　要：结合H.264压缩编码标准的特性,该文提出一种新的基于H.264/AVC的视频可逆脆弱水印算法。算法先计算当前宏块预测残差块量化的离散余弦变换(DCT)系数的哈希值生成认证码,再把认证码作为水印信息嵌入到下一个相邻宏块活性最大的4′4块的最后一个非零量化DCT系数中。在解码端,通过比较认证码和提取的水印信息进行视频数据完整性的认证。实验结果表明,该算法对视频质量和码率的影响较小,并且能对认证通过的视频数据进行还原和对认证失败的视频Ⅰ帧遭篡改区域进行有效的定位。Combining with the new features of the H.264/AVC video coding standard, a new reversible fragile video watermarking algorithm is proposed. The hash value of quantized Discrete Cosine Transform （DCT） coefficients ofpredictive residue blocks is performed with the proposed algorithm in current macroblock, and the authentication code is created. Then, the authentication code, as watermark is embedded into the last nonzero quantized DCTcoefficient in the most active 4 × 4 block of the next adjacent macroblock. The integrality authentication is carried out in the decoding end by the comparison of the authentication code with the extracted watermark information.The experimental results show that the proposed algorithm has little impact on the video quality and bit-rate, moreover, the original video can be recovered if the authentication is successful and the tampered area of I framecan effectively be located if the authentication fail.

关 键 词：视频水印 脆弱水印 可逆水印 H 264 视频认证 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]