基于矩阵Schur分解的彩色图像盲水印算法

针对目前大多数彩色图像数字水印算法中嵌入的水印是二值灰度图像且为提取水印时需要原始宿主图像或原始水印的现状,本文提出了一种将彩色水印嵌入到彩色宿主图像中的盲水印算法。该算法分析了矩阵Schur分解的基本特征,通过修改其上三角矩阵对角线上的最大能量元素来嵌入水印,水印提取过程达到盲提取的目的。实验结果表明,该水印方案不但嵌入和提取的速度快,而且具有较好的水印不可见性(峰值信噪比达到40dB以上)和较强的鲁棒性。

Color image blind watermarking algorithm based on Schur decomposition

In view of the present most of the color image digital watermarking techniques used binary or gray images as the watermarks and these techniques were non-blind extraction, this paper proposed a blind digital watermarking algorithm to embed a color watermark image into a color host image. The algorithm analyzed essential features of Schur decomposition, by modifying the largest eigenvalue element on the diagonal of the upper triangular matrix to embed the watermark. The watermark might be extracted from watermarked image without original host image and the original watermark image to achieve the goal of blind extraction. The experimental results show that the watermarking scheme has fast speed in embedding and extraction. In addition, the value of Peak Signal-to-Noise ratio more than 40 dB embodies the good invisibility. The solution also has a good robustness in terms of resistance of common image attack.