一种基于DCT和SVD的防篡改视频水印算法  

A Tampering Detection Video Watermarking Based on DCT and SVD

作　　者：袁世杰[1] 陆剑锋[1] 袁文强[1] 张善卿[1] YUAN Shijie;LU Jianfeng;YUAN Wenqiang;ZHANG Shanqing(School of Computer, Hangzhou Dianzi University, Hangzhou Zhejiang 310018, China)

机构地区：[1]杭州电子科技大学计算机学院,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2017年第1期51-56,共6页Journal of Hangzhou Dianzi University：Natural Sciences

基　　金：国家自然科学基金资助项目(61370218)

摘　　要：为了解决视频内容被篡改的问题,提出了一种防字幕篡改的数字视频水印算法.在算法中引入结构张量对字幕区域进行定位和处理,达到排除字幕背景干扰的目的.考虑到算法的鲁棒性,在水印嵌入过程引入奇异值分解(SVD).首先将视频解码到I帧的亮度(Y)通道;然后将其分块并进行离散余弦变换(DCT)、SVD分解;最后将水印信息嵌入到其中某些SVD分解系数上.实验结果表明,当文字的篡改面积大于15%时能检测出篡改字幕.在不可见性方面,每段视频嵌入水印前后的PSNR均超过44dB;在鲁棒性方面,受攻击后水印提取正确率在90%以上,具有着较好的不可见性和较强的鲁棒性.In this work,a digital video watermarking algorithm for tampering detection is proposed.The watermark is generated from the subtitle.However,the classical structure tensor is used for avoiding influence from the background of the frames in this work.The singular value decomposition(SVD)is introduced for the robustness in this work.Firstly,the I-frames in the video are decoded to the luminance(Y)frames.Secondly,the Y-frames are devised into the blanks in the discrete cosine transform(DCT)domain.Finally,the watermark is embedded the blanks that is after SVD.Experimental results demonstrate that the words that be tampered are detected.Furthermore,the invisibility and the robustness are strong.

关 键 词：水印 结构张量 离散余弦变换 奇异值分解 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]