基于NSCT及QR分解的零水印算法

Zero-watermarking algorithm based on NSCT and QR-decomposition

为有效对抗几何攻击,提出了一种利用图像几何校正的非抽样Contourlet变换(NSCT)及QR分解的数字图像零水印技术。该算法对图像进行NSCT变换,并在低频系数子图中提取稳定的SIFT特征点,根据特征点的变化求出几何变换参数。在对几何失真图像进行校正后,进行分块QR分解。取各块中R矩阵行向量2-范数组成序列,二值量化后变换为矩阵并与水印结合生成零水印。仿真实验结果表明,该算法能有效抵抗几何攻击,校正精度好,鲁棒性强,且能够避免水印鲁棒性与不可见性之间的矛盾。

In order to effectively resist geometric attacks, a digital image zero-watermarking algorithm based on non- subsampled contourlet transform （NSCT） and QR-decomposition is proposed. The algorithm applies NSCT to the image, extracts stable SIFT feature points from the low-frequency subblock, then figure out geometric parameters by changes of feature points＂ locations. When the watermark synchronous information is recovered, divide the image into blocks and apply QR decomposition in each block. Extract 2-Norm of matrix R in each block to form a sequence, then transform it into a matrix after binary quantization. The zero-watermarking can be formed by making the XOR operation between the watermark and this matrix. Experimental results show that the algorithm can effectively resist geometric attacks with high correction accuracy and strong robustness. By using the zero-watermarking method, it can also avoid the contradiction between the invisibility and the robustness.