Watermarking Technique Based on Harris-Laplace Feature Point Detector Capable of Resisting Geometric Attacks

Abstract

In this paper, a strong watermarking technique is proposed that first selects the more stable feature points of an image by using the Harris-Laplace detector and then adjusts the included angles among these feature points, centers of mass, and the most salient feature point to embed the secret message. Triangles are formed among the feature points before adjustment through Delaunay triangulation, and subsequently, the adjusted feature points are further connected. Finally, each triangle is fine-tuned through affine transformation to embed the watermarking. According to the experimental results, the proposed method possesses strong resistance to both geometric attacks and general image processing attacks.