Contourlet Domain Resist Geometric Attacks Watermarking Algorithm Research

With the rapid development of networks and information technology, the copyright protection of the digital multimedia products has become an urgent problem. As an effective means of copyright protection, digital watermarking technology has become a hot topic of research at home and abroad. The watermarking in the digital products may subject to a variety of attacks in the transmission process, and geometric attacks are very easy to make the embedded watermark fail to extract. Therefore, the anti-geometric attack watermarking algorithms are studied in this paper.Firstly, a robust watermarking algorithm in Contourlet domain based on singular value decomposition is put forward, for the invariant-domain anti-geometric algorithms having poor robustness against conventional attacks. The low frequency subband and the high frequency subband with the maximum energy obtained by Contourlet transform are selected as the target subbands, and then SVD are applied to them respectively. The watermarking information are embeded in the larger singular values. The invariance of singular values and the complementary between the low frequency and high frequency coefficients make the algorithm has good robustness against both conventional and geometric attacks.Secondly, a watermarking algorithm based on geometric correction using scale-invariant feature transform(SIFT) feature points matching is proposed, for the algorithms of feature points with smaller capacity having poor robustness against cutting. The low frequency subband coefficients after Contourlet transform are divided into blocks, and the watermarking information is embedded in all the singular values by quantifying the Euclidean norms of the singular value vectors. The watermark is embedded into the whole image, instead of relying on the feature points. By the advantage of the robustness of the SIFT feature point matching, a geometric correction method using the scale characteristics and spatial coordinates of the matched points is proposed. And before the watermark detection process, the geometric correction is applied to the image after attacks. The simulation results show that the algorithm can not only resistant to conventional attack, but also has good robustness for geometric attacks.