Robust Digital Watermarking Algorithms Based On Transform Domain And Optimization Strategy

How to prevent the illegal copy and unauthorized distribution of digital products is a difficult issue in the field of cyberspace security.Due to the high imperceptibility and high robustness of digital watermarking,its related technologies become one of the important means to protect the copyright of digital products.By analyzing the existing digital watermarking schemes,two robust digital watermarking schemes based on transform domain are designed in this dissertation,while the feasibility of these schemes is verified with simulation.The main contents of this dissertation are as follows.A digital watermarking scheme based on multiple transform domains and particle swarm optimization algorithm is proposed.Firstly,the host image is transformed by the lifting wavelet transform,and the high frequency component is further divided into 8×8 sub blocks.Then the discrete cosine transform is applied on each block and the intermediate frequency coefficients are extracted to form a new block.Then discrete fractional angle transform and singular value decomposition are performed on the intermediate frequency coefficient matrix,while Arnold transform and singular value decomposition are performed on the digital watermark.The digital watermark embedding algorithm is implemented by embedding the left eigenvalue vector of the decomposed digital watermark into the eigenvalue matrix of the transformed and decomposed host image.Besides,the particle swarm optimization algorithm is utilized to search for the optimal watermark embedding strength,in which the contradiction between imperceptibility and robustness can be balanced to some extent.In addition,the feasibility of optimization strategy is confirmed with contrast experiment and Simulation results demonstrate the robustness of the proposed watermarking scheme under various attacks.An adaptive blind digital watermarking scheme based on SVD block optimal selection strategy is devised.In this algorithm,the discrete wavelet transform is executed to the host image,and the low frequency component is divided into 4×4 sub blocks.Then the watermark embedding strength is obtained by calculating the information entropy of corresponding block,and the watermark is embedded into the location found out by SVD block optimal selection strategy.Simulation results show that this strategy can improve the PSNR value of the watermarking scheme up to 4.5 dB,and this watermarking scheme exhibits fine robustness to image compression,low pass filtering,motion blurring and other attacks.