基于小波变换和奇异值分解的图像水印算法研究

Image watermark algorithm base on the discrete wavelet transform and singular value decomposition

利用小波变换频谱特性和图像奇异值分解特征,提出了一种结合人眼对比感知特性的图像水印算法。并通过结合人眼视觉特性,将置乱的水印以一定的强度嵌入到图像的奇异值矩阵中,采用其逆过程提取水印,通过仿真进行了验证。对其实施了压缩、剪切、高斯噪声和中值滤波攻击测试,与前人提出的8种水印算法的抗攻击性能进行对比分析。结果表明,在质量因子为20%的较强压缩攻击下,提取水印的NC值仍能达到0.8359,含水印图的PSNR和SSIM达到25.0908dB和0.8451,且比8种水印算法具有更好的鲁棒性。综合表明,提出的算法有效地解决了水印嵌入过程中鲁棒性、视觉透明性与水印嵌入量之间的平衡问题。

Based on the spectral features of discrete wavelet transform and the singular value decomposition of image,an image watermarking algorithm combined with the contrast perception characteristics of human visual system（HVS）is proposed.By combining HVS characteristics,the scrambled watermarking are embed into the singular value matrix of image in some intensity,and the watermarking is extracted from the watermarked image using the inverse process of embedding watermarking.The algorithm is verified by simulation and is tested by four attacks,namely compression,shear,Gauss noise and median filter.The experimental results are compared with the ones of 8watermarking algorithms proposed by predecessors,respectively.The results show the NC of extracted watermark still can get to 0.8359 under the stronger compression attack whose quality factor only is 20%,the PSNR and SSIM of watermarked image can shill reach25.0908 dB and 0.8451.It has better robustness.The results show that the algorithm can effectively solve the problem to balance the relation among the robustness,visual transparency and embedding capacity of watermark.