Research Of Digital Watermarking Algorithms In Wavelet Domain Based On Chaos And Block-SVD

With the rapid development of information technology and computer networks, information security and copyright of digital product have become the practical problems and get widespread concern and research. Still digital image is used as the research object in paper, it combines chaos theory, lifting wavelet theory and singular value decomposition theory. Mainly research content is as follow.First, Algorithm one is "The research and realization of digital watermarking algorithm in wavelet domain based on double-Chaos". In the algorithm, in order to enhance encryption technology , preliminary encryption is processed by using Arnold scrambling on the binary watermark signal, producing a random sequences by using another key,. Last encryption was carried out on them. Wavelet coefficients of watermark information will be embedded in the low-frequency and high-frequency subband of the original image via wavelet transform. Adjust the depth of embedding through the result of experiment. The results shown that embedded watermark images have not only good invisibility but also have strong robustness.Second, Algorithm two is "The research and realization of digital watermarking algorithm in wavelet domain based on subimage extration and block singular value decomposition".The fourth chapter introuduces two kinds of watermark algorithm that are based on singular value decomposition. In the embedded stage of algorithm, subimages are extracted from the original image by chaotic sequence, then give the method to estimate embedded intensity or quantized value in order to realize watermarking embedding. The experimental results show that robustness and invisibility of digital watermarking algorithm have great balance.Third, Algorithm three is "A lifting wavelet for adaptive image blind watermarking algorithm based on double-chaos and block SVD." In fifth chapter have two algorithms, they are changed based on the algorithm two. It will change fix quantized value of algorithm two to revise the threshold by image of the local statistics characteristics adaptively, different threshold control coefficient rate strictly, the robustness of the two algorithms are superior to algorithm two.Finally, it's overall summarization of this paper and forecasting of further research work.