基于SVD和伪随机循环链的图像认证水印算法

SVD and Pseudorandom Circular Chain Based Watermarking for Image Authentication

针对矩阵的行列互换使得奇异值分解(SVD)后的奇异值不变,从而由奇异值直接产生水印安全性不高,以及基于SVD的图像认证水印算法把块水印嵌入本图像块的最低有效位,从而很难抵抗矢量量化攻击,提出了基于SVD和由Logistic混沌系统构成伪随机循环链的图像认证水印算法.首先通过Logistic混沌系统调制图像,使得图像矩阵唯一,然后将调制后图像块奇异值生成的块水印嵌入伪随机循环链对应图像块最低有效位.实验结果显示,该算法提高了水印的安全性,不仅能够准确定位,而且可以有效抵抗矢量量化攻击.

Elementary Transformation of Matrix doesn＇t change the singular value decomposition （SVD）, so the security of watermarks generated by SVD was not high. And most algorithms of watermarking for image authentication put watermarks into the least significant bits （LSB） plane of the block. These algorithms could not resist the counterfeiting attack. So a new watermarking algorithm for image authentication based on SVD and Pseudorandom Circular Chain （PCC） was presented. Firstly, the image was modulated by logistic mapping in order to make the image matrix unique. Then a Pseudorandom Circular Chain （PCC） was constructed by the logistic mapping. The watermarks generated by SVD were embedded into LSB of the corresponding block due to the PCC. The simulation results show that the proposed algorithm not only can localize the tampers correctly and accurately, but also can resist the counterfeiting attack.