Study On Watermarking Techniques Of Digital Images Based On Lifting Schemes

Digital watermark is a new direction in the information security field. It's a new technology which can protect copyright and authentication source. The author's copyright and personal information can be embedded in the products. But we can't percept it, only special software can pick it up. Digital watermarking technique is a effective method for copyright protection, which has become a hotspot in multimedia information security research and a core branch of hiding technology research. Wavelets have found wide applications in image processing due to its multiresolution and multiscale analysis property. In addition, the aspect of wavelets is helpful in managing a good distribution or location of the message in the image in terms of robustness versus invisibility. This paper takes still images as example, firstly introduces digital watermarking technology and its applications, then recommends present popular image processing technique in digital watermarking field—DWT. Additionally, algorithms of digital image watermark based on wavelet and error-correcting codes and on 3-band lifting wavelet transform are introduced. The main contents are described as followed:In this paper, the author firstly makes an introductory review on information hiding, digital watermarking technology and its applications. Then the author analyses fundamental principles of digital watermarking algorithm in detail such as the general framework, classification and the typical algorithm;An algorithm of digital image watermark based on wavelet and error-correcting codes is introduced in this paper: The algorithm is based on human visual system and the watermark is scrambled and encrypted by chaotic sequences to improve the capability of resisting geometrical attacking and security. Experimental results show that the proposed algorithm keeps the quality of the image well and is robust against common image processing, such as cut, JPEG compress, and noise;Additionally, a public algorithm based on 3-band lifting wavelet transform is given: The algorithm effectively improves robustness and the design of formula of watermark embedding primely balances the contradiction between transparence and robustness, moreover, it's a blind detection watermark since we detect the watermark without the original image;Thealgorithm is analyzed in detail, and experiment results again show that the distribution of image wavelet transform coefficients can be well modeled through the generalized Gaussian distribution. At the last, work of the thesis is summarized, and the further research direction is discussed.