Feature-based Watermarking using Discrete Orthogonal Hahn Moment Invariants

Many proposed image watermarking techniques are sensitive to geometric attacks, such as rotation, scaling, translation, or their composites. Even slight geometric distortions can also disable the watermark detector to reliably perform its function. In this paper, we utilize the robust invariant image features and discrete Hahn image moments to design a robust watermarking system that can withstand many geometric-distortions as well as survive a variety of common watermark attacks. Scale-invariant feature transform (SIFT) based bounding boxes and discrete orthogonal Hahn moment-invariants are used to embed watermark information into the selective image patches. Hahn moment-invariants are utilized for watermarking purpose because they are invariant to rotation, scaling, and translation. Watermark detection is performed by synchronizing the SIFT points and then computing the RMS threshold value between the original and the watermarked images. Several tests are performed to check the robustness of the proposed method. Experimental results validate the effectiveness of the scheme as well as prove that the proposed method is robust to several geometric attacks.