基于DWT-SVD域的彩色图像自适应水印算法

为了保护数字彩色图像版权,提出了一种结合离散小波变换(DWT)和矩阵奇异值分解(SVD)的彩色图像自适应水印算法。其主要设计思想是:先将原RGB彩色图像的各颜色分量进行小波分解;再将得到的中、低频带的小波系数与原水印图像分别进行奇异值分解,水印信息的奇异值重复嵌入到三通道中、低频带的小波系数奇异值中,且利用临界视觉阈值与奇异值之间的关系对水印嵌入强度作自适应调节,从而达到增强水印鲁棒性和确保水印透明性的目的。本算法适用于二值及灰度水印,在二值水印的提取过程中需设定一个提取阈值,以保证提取水印的完整性。仿真结果说明本算法对水印系统的透明性与鲁棒性作了较好的协调,是一种较稳健的算法。

DWT-SVD based adaptive color image watermarking

This paper proposed a color image adaptive watermarking algorithm combined DWT and SVD for copyright protection of digital color image. Firstly decomposing the color component of the original RGB color image using discrete wavelet transform respectively, it applied the SVD to HL, LL band and the watermarking alone, then embedded the same watermarking data into three channel by modifying the two band's singular values. The additional advantage of the proposed technique was taking advantage of the relationship between HVS and JND which could adaptively regulate the watermark embedding strength, so as to achieve the purpose of enhance the robustness and ensure transparency of watermark. The algorithm was suitable for binary and grayscale watermark, an extraction threshold value should be set in the binary watermark extraction process, to ensure the integrity of the extracted watermark. Simulation results show this algorithm is a more robust algorithm and better coordination the transparency and robustness of the watermarking system.