The Study Of Wavelet Packet Watermarking Based On DCT Image Elgenvector

With the rapid development of computer network and multimedia technology, information security and copyright protection of the digital works become practical problems need to overcome immediately.In the field of information security, digital watermarking is considered as a new method to protect the intellectual property rights. Considering security and robustness, transform domain watermarking is applied widely. In this paper, DCT-domain watermarking and DWT-domain watermarking of the transform domain are studied particularly.(1)As the digital watermarking technology and the basic theory of DWT are demonstrated, with the chaos theory and human visual system are introduced,two image watermarking algorithms based on DWT and wavelet-packet domain are designed. The performances of wavelet transform and wavelet packet transform are analyzed and compared in simulations.(2)Considering that the robustness of DCT-based and DWT-based image watermarking algorithm are different with the same attack, a watermarking algorithm based on combination of DWT and DCT in common way is presented.Experiment results show that the robustness of it improved remarkably.(3)For the traditional watermarking can't withstand geometric attacks well and the ones which can do it are somewhat complicated, a wavelet packet watermarking based on the DCT image eigenvector is proposed, exploiting the lobal DCT coefficients changing slightly when suffering the geometric attacks. The algorithm picks out part of the global DCT coefficients as a visual eigenvector of the image, combines it with the watermarking, then registers combination as a key to the third party. Experimental results show that the proposed method is effective.