Research On Image Digital Watermarking Based On Visual Model

Our society and people's lives have great changes due to information revolution. In information age, people are enjoying unprecedented convenience and happiness brought by explosive increase of digital media, however, at the same time they are also confronted with some new challenges. Especially, Digital media product copyright protection is an impending problem needed to be solved, but it cannot be solved effectively using Cryptology, Digital Signature and other traditional techniques. When people pursued some new ways, a new technique-Digital Watermarking emerged as the times require.Vision is a main means for people to observe the world. Nowadays, many research results on vision mechanism have effectively applied to image and video processing, and the result is very excellent. Visual model is to model HVS using mathematics method. In Digital Watermarking schemes, JND values can be calculated accurately using visual model, and accordingly embedded location and intensity can be decided appropriately in order to keep balance between imperception and robustness.This dissertation mainly focuses on research of still image Digital Watermarking schemes based on visual model. Visual models have been summarized in detail. Digital Watermarking schemes based on visual model have also been discussed roundly, and then a block classification watermarking scheme based on edge detection has been proposed with preferable robustness. Extended application of visual model on Digital Watermarking schemes is main work of this dissertation. A novel blind watermarking scheme using visual model to improve capacity has been proposed, and the experiment results show that the watermarking capacity has been obviously increased . A weighted Watermarking detector has been designed, and it can improve veracity of detecting Watermark. At last, Watermarking attack based on visual model has been discussed, and some feasible solutions have been brought forward.