Geometrically Invariant Watermarking Research Using Harris-Laplace Feature Point Detector

With the development of computer technology, the protection of digital information including integrity and copyright protection of digital information has been paid much attention. Digital watermarking technology is an effective way to solve this problem. It protects the security of confidential information, determines the copyright of digital product and detects whether the image has been illegally tampered. Currently, most of the digital watermarking schemes can resist common signal processing, but it is still facing a lot of problems for resisting geometric attacks. In many digital watermarking algorithms against geometric attacks, feature-based algorithms after a geometric attacks and common signal processing is still able to extract the watermark, so it has been widely used. But the feature-based algorithms is still has many problems, such as too much feature points, not stability, uneven distribution, the rotated image can not be effectively extracted and so on.To solve the above problems, this paper presents a strength and toughness of digital watermarking algorithm based on Harris-Laplace feature point detection. First of all, using the method of Harris-Laplace feature point detection to calculate the feature points of the original image, these feature points has the effect of scale invariant resist the geometric attacks. Second, compute the center of mass and maximum feature point to find the embedded position of watermarking, and embedded the watermarking sequence through the embedded method of angle and regional average value. Then, use the technique of Delaunay triangulation grid to connect the original feature points and the adjusted feature points, so they can form some triangle regions. Last, fine-tune the each triangle region by using the affine transformation to complete the process of embedding a digital watermark. In the process of embedding digital watermark, it can guarantee the correctness of extracting digital watermarking through the embedded method of angle and regional average value.In order to verify the effectiveness of the algorithm proposed in this paper, we do a simulation experiment by using the Matlab software. And comparing with Lee’s method, the algorithm proposed in this paper can well resist compression, rotation, scaling and other attacks and has good invisibility and robustness.