一种基于置乱和奇异值分解的鲁棒水印算法

A New Algorithm Based on Image Scrambling and Single-value Decomposition of Matrix

本文针对数字水印的抗攻击性问题,提出了一种新的基于置乱和奇异值分解的算法。算法首先将水印和载体图像同时进行置乱加密,再将置乱后的图像进行分块,然后对每块进行量化奇异值分解,在特征域嵌入水印。将双置乱技术引入分块量化奇异值分解方法中,进一步提高了奇异值分解算法的鲁棒性。其中,采用分块和量化特征矩阵的方法不用计算整个图像的SVD,大大缩短了水印嵌入和提取的时间。

This paper puts forward a new algorithm based on image scrambling and single-value deeomposition（SVD） of matrix. In this algorithm, the first step is to encrypt source image and watermark image by image scrambling technology. Then decompose the new source image in single-value by quantifying, and the watermark information is embedded into the diagonal matrix. It helps to enhance the robustness and reduce the time of embedding and extracting by using two kinds of image scrambling combine with the Block-Quantify-SVD.