The Research Of Blind Watermarking Based On Multi-fractal And Nonsubsampled Contourlet Domain

The digital communication network has developed rapidly, and the information release and transmission have become digitization and network by the Internet widely used. When people are enjoying their digital network communication to bring convenience, they also endure the misery brought by copyright confusion. It is very important to protect copyright in the digital communication network. In order to solve the problem, digital watermarking technology begins to walk into the sight of people.This report will begin with the background knowledge of digital watermarking and its previous research at home and abroad. With the hope of deeper understanding, the paper presents the definition of digital watermarking technology, including its basic concept, framework and embedded algorithm, such as time and space domain, DCT-domain and DWT-domain. This report put forward two kinds of improved embedding algorithm after studying the commonly used digital watermarking embedding algorithm. One is research on non-sampling of Contourlet domain transform digital watermarking research, which is based on pretreatment. non-sampling of Contourlet domain transform has good anisotropy, and produces little redundancy after embedding. At the same time, it also has the multi-resolution feature. So the non-sampling of Contourlet domain transform has its own irreplaceable advantages. And, its watermark is pretreated by Arnold. Experiments show that the algorithm has stronger robustness than conventional Contourlet transform algorithm. The other one is blind watermarking research on multi-fractal and non-sampling of Contourlet domain transform. The fractal images make good use of their self-similarity not only among each part but also between each part and entirety, which is suitable for the embedding of self-similarity and self-affinity watermarking images. Also, its fractal character improves the anti- attack capability of images. This research will apply Fractal Theory to digital watermarking so as to improve its anti-attack capability. Therefore, blind water marking research is put forward based on non-sampling of Contourlet domain and multi-fractal theory resulting from Fractal Theory. After the NSCT analysis of the original data and multi-fractal character analysis of the low frequency sub-band, the singular spectrum could be evaluated and this will result in a increased robustness if people embed watermark according to the singular spectrum. The experimental results show that this algorithm not only has the very good effect on embedding and extraction.,but also can extract high quality watermark under all kinds of attacks.