一种基于Harris—Laplace特征点检测的抗几何攻击数字水印算法

基于图像特征点检测的第2代数字水印算法具有良好的抗几何攻击能力．Harris算子是2代水印中使用较多的一种特征点检测方法．然而传统Harris算子提取的特征点对于较大尺度缩放攻击的稳定性较差，结合自适应尺度特性的Harris—Laplace算子在一定程度上克服了这一缺点，并且对于旋转，噪声及压缩攻击均具有较好鲁棒性．据此给出了一种基于Harris—Laplace特征点检测的数字水印算法，实验结果表明，该算法对于常规信号处理和几何攻击具有较好的鲁棒性．

A geometrically invariant digital watermarking algorithm based on Harris-Laplace feature points detector

The second generation digital watermarking algorithm based on image feature point detection has a good performance of robustness against geometric attacks. Harris detector is used widely in the second generation water- marking, but the traditional Harris detector is sensitive to image scaling. Harris-Laplace detector combined with scale adaption overcomes the disadvantages to some extent, besides, it has good robustnessagainst rotating, noise and compressing attack. According to this, a digital watermarking algorithm based on Harris-Laplace feature points detector is presented. The experimental results show that the proposed method is considerably robust against both geometric attacks and normal signal processing.