提出一种新的基于Contourlet变换的抗几何攻击水印算法。利用图像Contourlet系数在不同分辨率下的父子关系提出“空间方向重要树”(SDST)的概念,进而确定重要系数并嵌入水印信息;利用新的仿归一化方法,通过计算并比较原始图像(水印已嵌入)与接收图像(受到攻击后)的特征量校正图像,达到抵抗几何攻击的目的。结果表明,本文算法对翻转、尺度变换和旋转等攻击具有良好的鲁棒性,并且增加了水印信息的嵌入量,具有较好的不可见性。
This paper reports a new anti-geometric attack watermarking algorithm based on the Contour let transform. It presents the concept of space direction significant tree (SDST) using the Contourlet co- efficientsI father and son relationship at different resolutions, and determines the important factors to embed watermark information. This paper calculates and compares the eigenvalues between the original image (the watermark has been embedded) and the receive image (under attack) to correct the image by the new imitation normalization method for resisting geometric attacks. The results show that the algorithm has good robustness to flipping, scaling, rotation and other attacks. At the same time, it can ensure the watermark information robustness and imperceptibility.