一种新的鲁棒性水印算法

A New Robust Watermarking Algorithm

摘要: 本文提出了一种新的基于Radon变换、离散小波变换(DWT)与分块奇异值分解(SVD)相结合的数字图像水印算法。该算法利用Radon变换、小波变换和矩阵奇异值本身的特性，将原始图像进行Radon变换，然后将变换之后的图像作小波分解并将小波分解得到的高频子带进行分块，对每一块进行奇异值分解后，选取每块中最大的奇异值通过量化的方法嵌入水印信息。水印的检测和提取不需要原始图像。经过仿真实验证明，该方法对于旋转，缩放，平移等攻击具有很好鲁棒性的同时，对于普通的加噪，滤波，JPEG压缩攻击也具有很好的鲁棒性。

Abstract: This paper presents a new robust digital watermarking algorithm based on Radon transform, dis-crete wavelet transform (DWT) and block singular value decomposition (SVD). This algorithm uses the properties of the Radon transform, wavelet transform and singular value, and needs to be done in five steps: firstly, making the Radon transform of the original image; secondly, doing discrete wavelet transform (DWT) of the resulting image; thirdly, taking the high-frequency sub-band divided into blocks; fourthly, doing singular value decomposition (SVD) for every block; finally, selecting each largest singular value to embed the watermark information by the quantization method. The algorithm is a blind watermarking algorithm, because we can detect and extract the watermark information without the original image. The simulation results show that the method has good robustness not only for the rotation, scaling, translation and other attacks, but also for ordinary adding noise, filtering, and JPEG compression attack.