基于小波系数重要值与SVD的零水印算法

Zero-Watermark Algorithm Based on Important Wavelet Coefficient and SVD

为提高水印算法的抗攻击性能,提出了一种基于小波系数重要值与奇异值分解(SVD)的零水印算法.算法对宿主图像的低频子带进行分块,在密钥的控制下,伪随机地选择低频子带中的两块共4个系数组成一组,找出每组中的重要值;对重要值进行分块奇异值分解,最后通过判断每个子块最大奇异值的最高位数字奇偶性产生零水印.实验结果表明,该算法在抵抗普通攻击与抗旋转几何攻击方面均具有较好的鲁棒性.

In order to improve the performance against attack of watermark algorithm, this paper presents a zero-watermark algorithm based on important wavelet coefficient and Singular Value Decomposition（SVD）. Firstly, in a pseudorandom manner using a key, the original image is divided its low frequency band into blocks, two blocks of the low frequency band are chosed which form a group of 4 wavelet coefficients, then find out the important coefficients in each group. All important coefficients are divided into many blocks, each block conducts SVD. Lastly,zero-watermark is derived by judging the parity of the first digit of the biggest singular value in every block.Experimental results show the algorithm is robust to resist normal and rotation geometrical attack.