الفهرس 
Chapter?IntroductionOnly 14 pages are availabe for public view
 |  | from | 106 |  |  |
| | | from | 106 | | |
Abstract
Digital watermarking has been considered as an efficient solution for copyright protection of various multimedia contents. The bulk of the researches on digital watermark techniques have focused mainlv on classical media data types like audio, images and videos. However, extending these techniques to the newly emerged multimedia data type of 3D geometrical models is difficult as a result of the complexity encountered while handling the arbitrary topology and irregular sampling of 3D models. Moreover, the complex possible geometrical and topological attacks on watermarked model. This makes the 3D watermarking process more challenging.
The 3D Watermarking methods carefully hide some secret information bits in a 3D model. These methods contain contradictory parameters such as watermarked mesh quality (invisibility), capacity and watermark robustness. Therefore, these parameters must be formulated in objective functions and then trying to get their value v ia using an optimization technique. I hus this thesis proposes an optimized watermarking algorithm based on Non-Dominated Sorting Genetic Algorithm II (NSGA-II) optimizer. In this framework, the watermark is embedded into the low frequency coefficients in spectral domain using an embedding technique based on Dither Modulation (DM) technique. The goal of optimization is to search for optimal quantization step in order to improv e both quality of watermarked mesh and robustness of the w atermark. In addition, the proposed algorithm performance is analyzed in terms of signal to noise ratio and correlation coefficient. The simulation results show that the proposed algorithm can achiev e a good robustness against most of the attacks which were included in this stud.