| In recent years, with continuous development of the Internet and multimedia technology, digital image has become the main method to transmit and store the graphic information. Nonetheless, in the process of storage and transmission, image is always encountering duplication, manipulation and stealing and the traditional encryption method is incapable of satisfying the needs of image copyright protection. Thus, as an effective image copyright protection method, digital watermarking attracts more scholars’deep research. Compared to the method of direct embedding watermark in digital image pixel, digital watermarking, based on the transform domain, has higher robustness to the regular attack with better invisibility and larger capacity of embedding watermark. The method of constructing zero-watermark in transform domain, without changing the original information of host image and merely through the image feature value and watermark construction rules to construct watermark, effectively solves the contradiction between the invisibility of embedding watermark and robustness.This paper first introduces the research background of digital image watermarking and its research status at home and abroad then makes a detailed analysis of its basic framework and features, the theoretical foundation of transform domain algorithm and the evaluation standard for the digital watermark. On this basis, this paper, through combining the theories of encryption algorithm, makes a research on the zero-watermark algorithm in transform domain from the aspects of improving robustness of watermarking algorithm in digital image, strengthening the safety of image copyright protection and feasibility and practical application of watermarking algorithm. The main research and contribution of this paper are:(1) A zero-watermark algorithm based on Arnold for positioning the illegal manipulation.This algorithm transforms the host image by Arnold scrambling and makes exclusive-or operation on the least significant bit of the scrambling host and binary watermarking to construct zero-watermark and conceal the watermark information through chaotic mapping encryption. When carrier image is illegally manipulated, the manipulation position and shape will be defined by making absolute difference between the watermark tested and the original watermark information. The simulation experiment proves that stable space conversion rules scrambled by Arnold and large key space of chaotic mapping help to improve the safety of algorithm and algorithm has stronger robustness for common image treatment and attack. Lastly, this paper makes a detailed interpretation for the feasibility of this algorithm on medical application.(2) A resisting compression and geometric attacks zero-watermark algorithm in cellular automata domain.To aim at solving the problem of low robustness about general watermarking algorithm in resisting JPEG compression and geometric attacks, a zero-watermark algorithm based on cellular automata and singular value decomposition is proposed. This algorithm first converses the host image by two-dimensional cellular automaton and separates the low-frequency approximate subgraph and store the conversion rules of cellular automaton as the algorithm key; then divides the approximate subgraph into blocks and constructs zero-watermark through combining the singular value decomposition and cellular automaton in singular value matrix. Compared to the regular transform domain algorithm, cellular automaton has more conversion space and increases the safety of algorithm with key design in cellular automaton rules space; constructing zero-watermark with combination of cellular automaton conversion rules and singular value decomposition will increase the stability of algorithm. The experiment shows that this algorithm has better invisibility and safety with stronger robustness to the compressing and geometric transformation. |
PDF Full Text Request