基于 Krawtchouk 矩和 NSCT 变换的鲁棒零水印算法

摘要

为了解决现有数字水印中鲁棒性和不可感知性之间的矛盾，设计一种基于 Krawtchouk 矩和 NSCT（Non-subsampled Cont-ourlet Transform）变换的鲁棒零水印算法。算法首先对图像进行 NSCT 分解，其次计算其低频成分的 Krawtchouk 低阶矩不变量来构建特征向量，最后通过特征向量和水印的运算得到原始图像的版权信息。实验结果表明该方案对常见信号处理具有很强的鲁棒性，同时密钥的使用保障了算法的安全性。%In order to solve the contradiction between the imperceptibility and the robustness in current digital watermarking,we design a robust zero-watermarking algorithm which is based on Krawtchouk moments and nonsubsampled contourlet transform (NSCT).The algorithm first decomposes image with NSCT,then constructs the eigenvector by calculating the lower order Krawtchouk moments invariant of its low frequency component,finally obtains the copyright information of original image through the operation of eigenvector and watermark. Experimental results show that the scheme has strong robustness against common signal processing,meanwhile the security of the algorithm is guaranteed by using secret keys.