一种新的基于DCT和SVD的鲁棒水印算法  

Novel Robust Watermarking Algorithm Based on SVD and DCT

作　　者：王树梅[1] 赵向军[1] 

机构地区：[1]徐州师范大学计算机科学与技术学院,江苏徐州221116

出　　处：《小型微型计算机系统》2009年第5期967-970,共4页Journal of Chinese Computer Systems

基　　金：徐州师范大学自然科学基金项目(2008XLB16)资助

摘　　要：提出一种基于离散变换域(DCT)和矩阵奇异值分解(SVD)的鲁棒水印算法.首先将载体图像进行离散余弦变换,取包括直流系数在内的与水印图像大小一致的变换域信息;然后分别将所取信息和有意义水印图像作SVD分解,得到四个正交矩阵和两个对角矩阵;最后将水印图像SVD得到的三个矩阵分别嵌入到其余三个矩阵中,作SVD和DCT逆变换得到含水印图像.实验结果证明,本算法具有较强的鲁棒性.In this paper, a robust watermarking algorithm is proposed, which is based on a discrete domain transform（DCT） and matrix singular value decomposition （SVD）. First of all, the carrier image is processed by discrete cosine transform, and choosing some elements including the DC and its size equal to the significant watermark image;then, the SVD operation is respectively acted on the chosen information and the watermark image, four orthogonal and two diagonal matrices are received;finally, the three SVD matrices of the watermark image will be embedded in the three matrices of the carder. The watermarked image is get after inverse SVD and inverse DCT transform. The experimental results proved that this method is robust.

关 键 词：离散余弦变换 奇异值分解 数字水印 正交矩阵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]