The Design and Application of Adaptive Digital Watermarking Technology Based on the Perceptual Model
|School||Nanjing University of Information Engineering|
|Course||System Analysis and Integration|
|Keywords||digital watermark perceptual model adaptive human visual system (HVS)|
In the recent ten years the digital watermark technology receives more and more attention from all fields in the world. It has overcome the problem that content encrypted in the traditional cryptology technology is no longer safe after it has been deciphered and broken through. And it becomes an effective method in copyright protection of digital products and security of data on line.In this article, it describes the background, the status and the theories of the digital watermark technology at first. The article is about the copyright of digital image, so it describes the perceptual model, the human visual system (HVS) and the application of HVS in digital watermark technology. Then it introduces the mainstream algorithm of watermark imbedding at present, and it studies an adaptive watermark algorithm based on HVS. The born of the digital watermarking is the first step of the whole digital watermarking system. In this text we study the born of digital watermarking based on the edge characteristic of images. Then it splits the original image into 8~* 8 DCT blocks which not covered by each other. Computer the entropy and the nonzero number of each block, and according to the HVS, we can divide all blocks into four kinds of domain, including smooth domain, close-grained domain, fringe domain and the rest domain. The embedding strength varies with the features of the different domains. In the end we embed the watermark in different frequency coefficients according to the size of the original image and the born watermark, so that the watermarking can have a good robustness, invisibility and capability. This algorithm realizes the adaptive embedding of watermarking with the characteristic of HVS. It improves the security of the meaningful watermarking and the ability of anti-interference, and it will be more and more useful.We use a number of experiments to test the feasibility and accuracy of this algorithm, and the result of the experiments show the algorithm can resist several attacks such as addition of noise, cropping, median filtering and contrast enhancement.In this text, we design and realize an adaptive watermarking system based on the perceptual model to protect the copyright of digital images. A number of experiments indicate that this system behaves well in robustness, and it will be useful in application.