Robust Digital Watermarking Algorithm Resilient To Geometrical Transform

The development of multimedia and network brings forward a serious challenge to copyright protection of digital media. When encryption can not completely resolve the problem of copyright, digital watermarking as a new technology of copyright protection gets broad attentions in all over the world. Digital watermarking embeds hidden messages within digital media such as text, image, video and audio to protect copyright. The hidden messages may be copyright information or secret messages. At the same time these messages have a very little influence on host media and have specific restore methods. These messages should be invisible to illegal receiver.This dissertation introduced the development of digital watermarking roundly, and the current study and application in this field all over the world. We reviewed the watermarking theories in detail. In chapter 2, we introduced certain frequently-used watermark code technology. By these technologies, the security of the watermarking could be enhanced further. In chapter 3, the theoretical basis of watermarking on space domain and relative theories of the spread spectrum communication have been described. A method using the hopping technique is proposed to embed the image watermarking into DCT domain. According to the hopping pattern, the algorithm embeds the watermarking into the DCT domain repeated, which enhances the robustness of watermarking and keep it invisible. Experimental results show that the algorithm can stand against several attacks, such as geometrical attack, Low-pass filter, JPEG compression etc...In chapter 4, we introduced the theoretic basis of SVD (Singular Value Decomposition), and then give some mathematic analysis about the theory. In this chapter, two digital image watermarking methods based on SVD have been pointed out. One scheme is to embed the watermark into the original image's singular values firsthand, the other scheme is to embed the watermark into image's singular values after blocked the original image. Experimental results show that the two methods have obvious orientation of resisting attacks. Integrating the experiment result of the schemes in both the third chapter and the forth chapter, the following result could be discovered: The DCT has well robust in resisting the signal processing, while the SVD theory has sound mathematic theory basic in geometry distortions. A new robust watermarking algorithmbased on DCT-SVD is proposed in chapter 5. This is a blind digital watermarking, which can detect watermark without using the original image.