Research Of Image Watermarking Algorithms Based On Transform Domain

With the Internet’s explosive growth, the digital multimedia has developed rapidly. At the same time, a series of problems arises, such as the problem of intellectual property protection. In this context, the digital watermarking technology, as a new multimedia data for copyright protection of a new technology, has been rapid development and become a hot research topic in today’s academic circles all over the world since1990s. As a branch of digital multimedia, digital images are widely used and widely spread. With the rapid rise of e-ecommerce, the phenomenon of theft and tampering of digital images has become more and more serious, which not only violate the copyright of the original author, maybe also affects the economic interests of the original author. So it has become an urgent problem to protect the copyright of digital image.Considering the problems mentioned above.this article takes transform domain image watermarking as the main object of study. This thesis mainly studies the two aspects of DCT domain and DWT domain. The main contents of this thesis are as follows:(1) The conventional image watermarking algorithm cannot make compromises between watermark imperceptibility and robustness easily, to this situation, an optimal mean quantization blind watermarking algorithm was given. First to image watermark, carrying out Arnold scrambling to improve the security of the watermark, and then dividing the carrier image into8×8block of Discrete Cosine Transform, and according to the mean quantization principle,embedding the scrambed watermark in the smallest image with local quantization noise value, and the blind extraction can be achieved from the final watermark extraction. Experimental results show that the algorithm has strong robustness to many attacks such as adding noise,cut, filtering and so on, can also well get the watermark imperceptibility and robustness compromise.(2) In order to resist geometric attacks, an image watermarking algorithm based on the combination of DCT and SVD was given. Firstly, the image is block into8×8block of Discrete Cosine Transform. Based on the sensitive nature of the human visual system, selects the texture complexity sub-block. At last the encrypted watermark information is embedded into the complex texture of the image sub-block low-frequency coefficients. We can extract watermark information based on the key information of the watermark. The experimental results show that the algorithm has good watermark imperceptibility, and it not only has good robustness to common attacks, but also has good resistance to geometric attacks.(3) A color image watermarking algorithm based on DWT and SVD combination was given to protect color image copyright. Firstly, the alrorithm introduces the logistic chaotic scrambling encryption to improve the security of the watermark image.Then, the three color components R, G and B of carrier image are divided into4DWT decompositions, and then carried out singular value decomposition for the three components of the four low-frequency coefficients of wavelet decomposition. Finally, according to the human eye sensitivity to the type of R, G, B three color components, the watermark information is embedded proportionally singular value matrix of the three components, improve the imperceptibility of the algorithm. Experimental results show that the algorithm can resist various attacks such as adding noise, cut, JPEG compression and so on.