Research On 3D Model Watermark Embedding Method

With the development of network technology, more and more digital products come into public network. Meanwhile, the copyrights protection problem becomes crucial. By embedding some tags into multimedia contents, digital watermarking techniques can protect their copyrights. In recent years, with the development of CAD, biomedical imaging, 3D animation and other applications, the copyrights protection for 3D models become necessary. Therefore, 3D model watermarking technique has been widely researched.In this thesis, we would like to discuss some key problems in 3D model watermarking technique, and research as a proprietor and as an attacker on the robustness, localization, feature selection and data embedding problems.The main contributions include:(1) A semi-fragile watermarking algorithm for 3D models based on integral invariants is proposed. As the robustness is not well defined for 3D model watermarking, we list some commonly-used attacks and class them into malicious attacks and non-malicious attacks. Based on integral invariants, this algorithm can survive under similar transforms, and be robust against noise and cropping.(2) An optimized framework for indexed localization scheme is proposed. To improve the capacity and invisibility of 3D watermarking algorithms, we optimize the watermarking scheme. By solving a maximum independent set problem and an assignment problem, the overall performance of the algorithms are greatly improved.(3) A new counterfeit original attack method for 3D model watermarking algorithms is proposed. As an attacker, we check the validity of existing algorithms. By reordering the vertices, the proposed method can counterfeit a fake original model, scramble the extracted watermark and confuse the authentication decision. With an example, we show the feasibility, succeeding conditions and method to avoid.