| Nowadays in a society with information, information has become a kind of momentous resources, information security has also become a main battlefield of international competition in 21st century. In order to protect their own economic and politic interests, every government attaches importance to network and information security, information security has become a global problem. In a network society the protection of multimedia content has become a crucial issue that attracted much attention of both the government and the public.As a main carrier of multimedia information, digital image has characteristics of copy and tamper easily, the copyright protection of digital image and the authentication to verity of digital image have become the two most crucial issues to be solved. The methods of digital image robust watermarking and fragile watermarking have provided an approach to resolve the problem issued above. Because that there is a natural relationship between chaos and information encryption. Herein, this dissertation deals with multipurpose digital image watermarking, our main works are as follows:At first, we probe into the basic principle of watermarking technology, which involve the classification of watermarking, principle of image watermarking, algorithm designing for watermark, attacks and evaluations to watermarking and the application of watermarking. We probe into, as well, chaos and its applications in watermarking. Those applications involve mainly the concept of chaos and methods to produce chaos sequence and its property; the methods to produce {0,1} sequence by using chaos sequence, the chaotic system can be used to scramble and encrypt watermarks. We have provided an example to modulate watermarks by using chaos sequence.Secondly, we propose an algorithm based on chaos to achieve the aim of copyright protection and content authentication. In order to achieve the aim of copyright protection, we use an image having meaning as invisible watermarks. After being scrambled, the invisible watermarks are embedded to the most frequency AC coefficients of DCT domain of image block. In order to achieve the aim of content authentication The semi-fragile watermarks produced by chaotic system are embedded into the second low AC coefficients of DCT domain of image block.. Both of them can be extracted without the original image. Simulation results demonstrate the effectiveness of our algorithm in terms of robustness and fragility. The robust watermarking is robust to mild modification such as JPEG compression, channel noise and general mild image processes. In contrast with robust watermarking, the fragile watermarking is fragile to malicious attacks and it can locate the attack position.Thirdly, we propose an algorithm based on chaos to achieve the aim of image copyright notification and copyright protection. We take, especially, the safety of watermarking into account. In this dissertation, we probe into the embedding technique of visible watermarking and propose a method of generating coefficients for visible watermarks embedding in spatial domain by using 3-D Liu chaotic system. We also discuss the characteristics of DWT, at the same time, new techniques are proposed to utilize image protection watermarking to embed image block mid-frequency coefficient of WT. General two-dimension Arnold transform is utilized to scramble the watermarks' embedding positions. The simulation demonstrates that the algorithm is effective. The algorithm can achieve both copyright protection and copyright notification. Legal users can remove the visible watermarks, simultaneously, illegal users can neither remove the visible watermarks nor extract the invisible watermarks. The invisible watermarks can resist most attacks. Moreover, the quality of picture after removing visible watermarks is high, namely, the invisible watermarks have a little influence to the host image. The algorithm shows that it is robust enough against many common attacks.At last, the dissertation points out some problems existed in digital watermarking research, and highlight some further research directions. |
