Research On Blind Robust Watermarking Scheme Of Digital Color Image

With the rapid development of communication technology and network technology,the digital image have been widely spread on the Internet,copyright infringement and information tampering of which are occurring more frequently.Robust imge watermarking technology which is the most effective method on the issue of copyright protection of image has attracted more and more attention at present.Most of the existing robust image watermarking algorithms were mainly designed to mark grayscale images,however,a few schemes were proposed to color image.The research on digital watermarking of this paper puts emphasis on the design of robust color image watermarking algorithm.Many existing works can not meet the requirements of the robustness and invisibility of the watermark performance.As for the existing watermarking algorithm based on Contourlet transform failing to extract the watermark information completely without attack,a improved blind color image watermarking algorithm based on Contourlet transform and Hessenberg decomposition is proposed in this papper.Firstly,each color channel of the host image is transformed by Contourlet transform and its low frequency subband is divided into 4x4 non-overlap coefficient block.Secondly,the coefficient block is selected by pseudo-random algorithm.Thirdly,the permuted watermark information of color image is embedded into the coefficient by quantified the sum of block elements,and the watermark is re-embeded into the coefficient block if there is data overflows.Experimental results show that the proposed scheme is robust against most common attacks including JPEG lossy compression,JPEG2000 compression,median-pass filtering,rotation,scaling,blurring and noise addition et al.Many existing color image watermarking schemes perform passablely at the expense of high computation complexity.In this study,a improved color image watermarking algorithm based on the Hessenberg decomposition.Firstly,each color channel of the host image is divided into 4×4 non-overlap blocks.Secondly,the blocks selected by pseudo-random algorithm is processed by texture treatment.Thirdly,block is decomposed by Hessenberg decomposition and the second row second column element in the Hessenberg matrix H is quantified for embedding the permuted watermark information.Extensive simulation results and comparisons show that the proposed scheme has higher imperceptibility and robustness against most common image attacks in comparison with other related methods.Especially,the proposed method has lower computational complexity than other methods based on Schur decomposition or singular value decomposition.