Research On The Algorithms Of 3D Digital Watermarking

With the development and application of digital multimedia technology and its wide distribution on Internet, it's more and more easy for people to get these products of digital multimedia. But, the convenience of Internet has also prompt unauthorized occupancy, duplication, modification and distribution of valuable contents. The copyright protection of digital multimedia is a important issue of common concern to academic and industrial institutions. As a new scheme of copyright protection of digital multimedia, the digital watermarking receive a lot of attention. 3D-modle, a new kind of multimedia data, it become more and more widely used. 3D digital watermarking has became a hot spot in the research field of digital watermarking. This dissertation researches and discusses the techniques of 3D-model watermarking for 3D mesh model. The main works we made are as follows:1)This dissertation first introduced the background and significance of research on digital watermarking, basic concept, characteristic, classification, the field of application and the development of 3D digital watermarking at home and abroad. Then the concept, the principle of system, the characteristic and performance evaluation of 3D digital watermarking were introduced.2)After the concept of frequency-domain watermarking and feature point were introduced, a frequency-domain watermarking scheme based on feature point was proposed. The main idea is that we pick out the feature points form high variation areas of the 3D model. We made a norm sequence of norms of features points, and group the feature points into different sections according to their norms. The sections are transformed into signal in frequency domain and embed the watermark into the signal.3)The concept and some classical algorithms of 3D blind watermarking were introduced and the defect of a algorithm was analyzed, and proposed a better algorithm. The main idea is that we pick out the feature points form high variation areas of the 3D model. Then we compare the norms of feature points with the mean of norms of neighbor points, modify the norms of feature points according to the watermarking.