Robust Watermarking Against Geometrical Attacks

With the great development of multimedia technology and the popularity of Internet,distributing of digital products through Internet is more and more convenient. Meanwhile,the problem of multimedia copyright protection becomes increasingly serious. In recentyears, robust digital watermarking technology has become an effective way of copyrightprotection. Because geometric attacks have greatly affect the robustness of watermarkingsystem, so the research on the counter-geometric transform attacks is considered as adifficult and hot problem to improve the robustness of watermarking algorithm.The paper is focused on the digital watermarking system resilient to geometric attacks,the main research results are listed as follows:(1) The research status of watermarking algorithms resilient to geometric attacks wasparticularly analyzed and reviewed, they were divided into two kinds to summarize, onewas based on the invariant, the other was the attacks revised.(2) To improve the security of digital watermarking technology, the method of Arnoldwas analyzed and realized by the programming tool.(3) The existing image watermarking algorithms using singular value decomposition(SVD) was simulated and its flaw was pointed out, then a new image watermarking systembased on Disperse Fourier Transform and SVD is proposed. This scheme sampled thecoefficients from the amplitude of frequency for being recomposed as a square matrix, thenpartitioned the matrix into blocks and performed SVD transformation. The watermarkingwas embedded into the largest coefficient from each component. The experiments provedthe validity and the robustness of the scheme.(4) Based on the image geometric moment to counter-geometric attacks was proposed,this scheme used the information of geometric moment of the watermarked image to revisethe attacked watermarking cover, and achieved the robustness of semi-blind detectionagainst signal processing attacks and geometric attacks. The watermark was embedded intothe low frequency of wavelet by using a predefined quantization coefficient. In addition,calculating an invariant centroid of the image was used to embed the watermarking.