Researches On Watermarking Algorithm And Its Application To Image Software Of Electronic Signature

With the development of digital and network technology, various forms of multimedia digital works have been applied more and more widely in the network. However, there are also important problems to solved, such as the convenience and insecurity. Digital watermarking technology is an effective copyright protection and security maintenance technology. Digital watermarking has got more and more attention in multimedia security area in recent years. The research on robust digital watermarking algorithm is still a hot topic, especially the geometrical resistant watermarking algorithm. To improve the robustness of digital watermarking algorithm, this thesis focuses on three stages of a digital watermarking system: research of the generation, embedding and extraction/detection of the watermark is studied. And two watermarking algorithms are proposed. And these algorithms are applied to the electronic signature systems, which can guarantee the security of seal image. Main contents of this thesis are as follows.ⅰ. The image software of electronic signature is designed, which can provide different shapes and functions according to various requirements of users, and with this image software, users can save several kinds of seal images.ⅱ. Traditional digital watermarking information usually was generated by the gray image, more embedded information result in largely modifying more image data, as well as the influencing the invisible of watermarking images. A bit-decomposition watermarking algorithm based on the mask is proposed. Experimental results show that the proposed scheme not only has better visual quality, but also is very robust against common image processing and geometric attacks.ⅲ. To chose the stability and hidden of embedding position, a geometrically synchronous watermarking algorithm based on the corner feature is proposed. Sufficient experiments show the robustness of this approach, especially against resisting geometry attack such as Rotation, Scaling and Translation.ⅳ. To ensure the security of seal image in electronic signature systems, our approach combines both the fingerprint recognition and digital watermarking. The watermark signal is generated by fingerprint feature information, and hidden in the host image. Experimental results show that by using our approach, robustness of seal image is much better, which can increase the security of electronic signature systems.