ResearcResearch Of Digital Watermarking Technology Based On SVD

Singular value decomposition(SVD)is one kind of orthogonal transformation,it can diagonalize matrix. Its dominancy lies on singular values which embodyrelatively stability. By virtue of this characteristic, people utilize the mathematicaltheory among digital watermarking and acquire nice effect of watermark embedding.However, based on the phenomenon that singular value decomposition is failure toextract in digital image and the diagonal distorts, people combine this theory withothers further and propose the new digital watermarking algorithm.Based on singular value decomposition theory, the author proposes the newdigital watermarking algorithm by combining it with discrete wavelet transform andzero-watermark technology. We draw the conclusions as follows:(1)With the color images becoming the main medium for information communicating,people proposed the color image digital watermarking. While the existing methods are embedding,the watermark images are always the binary images or gray scale images. The thesis is directed tothe demand which the color images are treated as watermarking, the author has the mind thatthe mass of data produced when the color images embed regarded as watermark,extracting three RGB components of watermark to embed to three RGB componentsof carrier images, furthermore, combining with Discrete Wavelet Transform(DWT)in this process. The algorithm sorts out the problem that the mass of data producedwhen the color images embed regarded as watermark, and it embodies specialanti-attack capability through the experimental results show.(2)In order to solve the contradiction between invisibility and robustness ofthe watermark, people extract the zero-watermark algorithm. Nevertheless, most ofthe algorithms are all complex nowadays and the zero-watermark combined withsingular value decomposition always adopt bigger singular value after itsdecomposition to embed into watermark. In allusion to this issue, the author reaches aconclusion that the least singular value is more relatively stable than the biggest one,through linking the singular value decomposition to the zero-watermark technology,raises the new algorithm and citifies the stability of the least singular value viaexperimental data.