基于NSCT-DCT-DWT-SVD联合数字水印算法  

JOINT DIGITAL WATERMARKING ALGORITHM BASED ON NSCT-DCT-DWT-SVD

作　　者：展虎 毛力[1] Zhan Hu;Mao Li(School of Internet of Things,Jiangnan University,Wuxi 214122,Jiangsu,China)

机构地区：[1]江南大学物联网工程学院,江苏无锡214122

出　　处：《计算机应用与软件》2018年第3期309-315,共7页Computer Applications and Software

基　　金："十二五"农村领域国家科技计划子课题(2015BAD17B02-8);现代农业产业技术体系专项资金项目(CARS-49);江苏省产学研合作项目(BY2015019-30);江苏省自然科学基金项目(BK20140166)

摘　　要：针对现有已提出的相关数字水印算法在不同攻击下存在的问题,提出一种将非下采样Contourlet变换NSCT(Non-subsampled Contourlet Transform)和DCT-DWT-SVD联合的数字水印算法。该算法将原始的宿主图像进行非下采样Contourlet变换得到与原始图像大小相同的低频子带,将低频子带进行离散小波变换分解,并对小波分解的低频部分与Arnold变换后的水印进行离散余弦变换。将变换得到的新区域与变换后的水印信息图像进行奇异值分解,将两者分解得到的奇异值矩阵相结合,作为最终的水印嵌入主成分。实验结果表明,该算法可以有效地抵抗JEPG压缩﹑椒盐噪声﹑泊松噪声﹑滤波等不同类型的攻击,在确保水印不可见性的同时具有较高的原始图像保真性。A digital watermarking algorithm based on joint Non-subsampled Contourlet Transform(NSCT)and DCT-DWT-SVD is proposed for solving the problem of attacks toward the existing digital watermarking algorithms.The algorithm performs the Non-subsampled Contourlet Transform(NSCT)on the original host image to obtain the low frequency sub-band with the same size as that of the original image.Decompose the low-frequency sub-band by using Discrete Wavelet Transform(DWT)and conduct the Discrete Cosine Transform(DCT)on the low-frequency part of the wavelet decomposition and the watermark which is transformed by Arnold.We apply the Singular Value Decomposition(SVD)on the obtained area and the watermark,and combine the two singular value matrices that obtained by the Singular Value Decomposition as the final watermark embedded principal component.The experiment shows that the algorithm can resist the different types of attacks,such as JEPG compression,impulse noise,Poisson noise and filters effectively.It ensures the invisibility of watermark with high fidelity of the original image.

关 键 词：数字水印 非下采样CONTOURLET变换 DCT-DWT-SVD 水印嵌入主成分 不可见性 保真性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]