The Application Of XML Technology In Digital Watermarking

The rapid development and the popularity of internet have accelerated the arrival of digital age, which are providing the great convenience for multimedia information transmission and processing, make the expression of information more efficiently and accurately. However, severe problems have emerged obviously regarding copyright protection and falsification of digital products. It has been more difficulty to track the piracy. Digital multimedia information security problem has become one of the important and challenging hot research topics. In this back ground, digital watermarking as a new technology for copyright protection has got high attention by the domestic and international academia as well as the business community. Based on the existing survival techniques, this paper considers the static digital image as the main research object, combined with fractal theory, mainly concentrates on the application of XML in digital watermarking. The main contributions are as follows:Firstly, using the self-similarity of the fractal image by researching the principle of LS fractal model, and making the XML document store the information of triple G= based on the tree structure of XML, and then implements the transform algorithm between fractal images and XML document according to the design and specific implementation steps illustrated. Both examples and experiments indicate that use XML to represent the fractal image not only have readability but also decrease the storage space of fractal image greatly, which make the process and display of fractal images more flexible and quick. Thus, the watermark is meaningful images rather than 0,1 code, that make the watermark more persuadable as the copyright identification.Secondly, as the shortcoming of the existing watermarking technology based on fractal theory, the algorithm proposed in this paper is combined XML technology with fractal theory, using XML to represent the fractal image and embed the XML into the host image. In addition, CIE L*a*b* color space is used in the watermark embedding and extraction process, making the watermarked image more imperceptible to human eyes. Simulated experiments prove the proposed method has significantly improved imperceptibility and robustness compared to the traditional method, but because the watermarking algorithms based on spatial domain, the robustness under JPEG compression attacks is week.Thirdly, since the existing watermarking technology based on DCT leads to lower efficiency due to the lack of consideration of the feature that different color images have different color distributions, choosing the fixed color component to embed and extract the watermark. To improve the traditional method, a new watermarking algorithms based on Watson visual model is proposed with the use of the masking characteristic in color images. The scheme carries out the DCT transform in each color component and then chooses the appropriate color component dynamically to embed and extract the watermark in the DCT intermediate frequency coefficient. Experimental results and analysis show that this method has better effect in imperceptibility than the existing ones, and better robust for various attacks. Furthermore, the scheme solves the problem of small capacity in the watermarking algorithms base on DCT by using XML to represent the watermark.