The Research Of Robust Image Watermarking Method Against Geometric Attacks

As a traditional encryption technology effective complement means, Digital watermarking is the protection of copyright or authentication source and the integrity under the opening network a kind of new technology. In recent years, It has aroused people's attention and has become a hot spot of international academic research. As a new technology, Watermark has the underlying research value, also have transparency and robustness of the basic demand. So robust is a basic requirement of the digital watermark system and the main research contents in recent years. This paper focus on the digital image watermarking domain to resist geometric attacks problem is studied and discussed, the main research is as follows :1.Based on the Support Vector Regression and the Tchebichef moments, we Propose a robust image watermarking algorithm in Nonsubsampled Contourlet Transform(NSCT) domain. Firstly, the NSCT is performed on original host image. Then, the selected low-pass subband is divided into small blocks. Finally, the digital watermark is embedded into original image used repetition strategy by modulating the NSCT coefficients in each small block. The detecting procedure include: Compute some low-order Tchebichef moments of test image and input the model SVR which has been trained to gain geometry correction parameter; Finally, the test image is corrected. Extract watermark in each small block and statistical results. Because SVR has very good learning ability and the Tchebichef moment has the highly accuracy, Make in different attacks watermarking all can be accurately extracted, thereby improving watermarking robustness.2.Using image affine covariant features, we Propose a new digital image watermarking scheme based on local visual attention. Firstly, The SIFT algorithm is used to extract feature points of a given host image. A set of stable and non-overlapped ellipse affine covariant regions are then chosen via selection criterion based on the minimum spanning tree clustering algorithm. Then, ellipse normalization is performed on these circular feature regions to achieve scaling and rotation invariance. The Nonsubsampled Contourlet Transform(NSCT) is performed on these circular regions, and finally embed the watermark in the transformed low-pass subband. The blind detection algorithm is presented. Simulation results demonstrate that the proposed scheme is robust against common image processing operations, geometric attacks, and combined attacks.