Multiple Robust Digital Watermarking Algorithm Based On DWT And SVD

Information sharing based on computer network provides a convenient way for the useand transmission of multimedia digital works. However, the information security protectionand copyright protection of digital works become the practical problems which urgently needto solve, for the digital works can easily be illegally copied and tampered with. In order toprevent digital works from infringement, piracy and malicious tampering, digitalwatermarking technology combined with the authentication and identification features oftraditional cryptography is valued by more and more people, and becomes a very popularfield of study. Digital watermarking technology can provide effectively protection to thecopyright of digital works by embedding the copyright information in multimedia data. As animportant branch of information security research field, digital watermarking is an effectivemeans for copyright protection of digital products, and it has a wide range of researchprospects.This paper gives an overview of the basic theory of digital watermarking technology,common attacks, performance evaluation, as well as the typical algorithms of digitalwatermarking, etc. Then the paper had a introduction on the watermark image preprocessingtechnique, including the watermark scrambling technology and bit-plane decompositiontechnique, etc. And then a robust multiple digital watermarking algorithm based on DWT andSVD is proposed, which combined with the characteristics of robustness to geometricdistortion of singular value decomposition and the characteristics of robustness to generalattack of discrete wavelet transform. Based on the bit-plane decomposition technique, thealgorithm makes the gray-scale watermark with rich visual information can be quantitativelyembedded. In this paper, watermark is embedded in the wavelet domain of the original carrierimage, namely watermark is quantitatively embedded in the transform coefficient obtained byusing block cosine transform in the low frequency sub-band, at the same time watermark isquantitatively embedded in the singular value vector obtained by using block singular valuedecomposition in the high frequency sub-band, and zero watermark is extracted for copyright protection of digital images according to the characteristic information of the intermediatefrequency sub-band. Finally, in order to improve the robustness of algorithm to resistgeometric attacks, the same watermark information is quantitatively embedded in the airspaceof the carrier image. The simulation results show that the algorithm can effectively resistvarious attacks and has good robustness and transparency.