Research And Implementation On Digital Image Watermarking Technology Against Geometric Attacks

With the rapid development of multimedia technology and the popularity of internet, distribution of digital products through internet becomes an important way of new civilization, how to protect the author's copyright becomes an important research topic. Digital watermarking technology is an effective way of copyright protection in recent years. At present, the ability of counteracting geometric attacks hasn't completely achieved the request of application. The geometry transform attack is considered as a pivotal problem to improve the robustness of watermarking algorithm. Considering the problems in information security and copy protection, this dissertation makes a wide and deep research on the current hotspot in digital watermarking technology, such as the image watermarking technology against geometric attacks.The main research contents of this dissertation are recapitulated as follows:1. The fundamental concept, general model, research situation and development trend of digital watermarking are introduced. Then the present situation, character and shortage of geometric attacks invariant watermarking algorithm and watermarking algorithm based on discrete wavelet transform (DWT) are analyzed.2. The image moment and Zernike moment which were widely applied in pattern recognition are introduced, and the geometry transform parameters estimate algorithms based on the image moment are researched and implemented. These two algorithms all can be implemented with less image moment information to get the geometry transform parameters, the Zernike moment has the character of rotation invariant which is more widely applied in rotation invariant watermarking technology.3. A geometric attacks invariant wavelet field digital watermarking algorithm based on image singular value decomposition (SVD) and Zernike moment is designed. In this algorithm, the singular value of the four sub-bands which are decomposed form the original image are modified with the singular values of the watermarking, then the destroyed image can be renewed from the geometry transform parameters estimated by Zernike moment. Experimental results show that this algorithm has better invisibility and robustness and can exactly estimate the geometry transform parameters of rotation and scaling. And this algorithm also has better robustness to the union geometric attacks which regards the rotation attacks as the main part.