融合DCT与SVD的半脆弱图像水印算法研究  

Semi Fragile Image Watermarking Algorithm Based on DCT and SVD

作　　者：饶俊慧 吴晓云 RAO Jun-hui;WU Xiao-yun(School of Physics and Telecommunication Engineering,YuLin Normal University,Yulin Guangxi 537000,China;Information Science School,Guangdong University of Finance&Economics,Guangzhou Guangdong 510320,China)

机构地区：[1]玉林师范学院物理与电信工程学院,广西玉林537000 [2]广东财经大学信息学院,广东广州510320

出　　处：《计算机仿真》2022年第7期507-511,共5页Computer Simulation

摘　　要：半脆弱图像水印算法能够对图像的完整性进行认证,针对现有算法在攻击时表现出的鲁棒性不佳问题,提出了融合DCT与SVD的半脆弱图像水印算法。为了增强水印的隐藏性,利用奇异值具有的稳定性和对主观效果的非影响性,设计了基于独立分块低频子带的SVD分解,在RGB通道内分别嵌入半脆弱水印,并根据PSNR与NC对嵌入系数做出了限定。SVD处理的同时,利用二值水印感知分块与分组的篡改攻击。为了更好的感知水印,融合二维DCT变换增强水印的嵌入与提取。基于MATLAB与VOCdevkit的仿真,得出算法的PSNR指标始终保持在40dB以上,NC指标始终保持对1的良好趋近,证明融合与的半脆弱图像水印算法能够有效保证水印嵌入与提取的误差,具有良好的非可见性,即使在各类攻击情况下,仍能保证良好的鲁棒性。The semi-fragile image watermarking algorithm can authenticate the integrity of the image.Aiming at the poor robustness of the existing algorithms,a semi-fragile image watermarking algorithm based on DCT and SVD is proposed.In order to enhance the hiding ability of the watermark,the stability of singular value and its non-influence on subjective effect are used.The SVD decomposition based on an independent block low-frequency sub-band is designed.The semi-fragile watermark is embedded in the RGB channel,and the embedding coefficient is limited according to PSNR and NC.At the same time,the binary watermark is used to detect the tamper attack of block and group.In order to better perceive the watermark,a two-dimensional DCT transform is fused to enhance the watermark embedding and extraction.Based on the simulation experiment of MATLAB and VOCdevkit,it is concluded that the PSNR index of the algorithm is always above 40 dB,and the NC index is always close to 1.It is proved that the semi-fragile image watermarking algorithm combining DCT and SVD can effectively guarantee the error of embedding and extracting watermark,and has good invisibility.Even in all kinds of attacks,it can still guarantee good robustness.

关 键 词：图像水印算法 奇异值 独立分块 低频子带 离散余弦 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]