基于SIFT特征的公钥水印算法

Public-key Watermark Algorithm Based on SIFT Feature

针对现有公钥水印算法抗几何攻击能力弱的问题,提出一种利用SIFT特征实现抗RST攻击的公钥水印算法,采用SIFT算法对原始图像的特征点进行提取,构造局部特征区域,并在其DCT域中嵌入水印信息,使嵌入水印后的图像能够更好抵御RST攻击,从而满足公钥水印可公开检测的要求。实验结果表明,该算法能够有效抵抗RST攻击和一般信号处理的攻击。

Aiming at the weak robust to the geometric attacks of the public-key watermark algorithm, a public-key watermark algorithm based on Scale Invariant Feature Transform（SIFT） is proposed for resisting Rotation Scale Translation（RST） attack in this paper. The feature keypoints of original image are extracted. The watermark in the DCT domain of the feature zone is embedded, so that it can resist the RST attack better, and the verification of the watermark is public. Experimental results demonstrate this algorithm can resist general RST attacks and signal processing attacks.