Research On The Digital Watermarking Algorithms Based On Wavelet Domain

With the rapid development boom of internet and digitization of multimedia resources, the problems of infringement, pirated illegal copy and malicious tampering of digital works are getting more serious, which makes digital watermarking an important research focus.Firstly, the dissertation introduces the basic principle and characteristics of the digital watermarking as well as its research status domestic and abroad, and then gives the basic framework of every parts.compares the strengths and weaknesses of the existing watermarking algorithms and analyzed some existing questions of current digital watermark technology.In the watermark embedding algorithm, this paper focuses on DWT transform domain digital watermarking embedding technique, which has been widely used as data compression, processing and analyzing tools in digital watermarking because the DWT transform domain possesses advantageous of space domain and DCT transform domain. In this dissertation, two kinds of transform domain watermarking algorithm has been proposed, and a machine learning method named supports vector machines also has been introduceed to maximize the overall performance improvement of the watermark.The mainly research result is as follows:1. A digital image watermarking algorithm in wavelet transform domain is proposed. With this algorithm, combined the watermark scrambling transformation and human visual system characteristics, the watermark information after pretreatment are embedded into middle frequency subband of the images under different scales of wavelet transform. The original image is required to take part in extraction, whose algorithm belongs to non-blind watermark.2. A digital image watermarking algorithm based on the support vector machines is proposed. The watermarking schemes are proposed based on the coefficient direction-tree structure in the discrete wavelet transform domain. Firstly, the concept of wavelet coefficient direction-tree is defined. then the inherent relation between root node and its offspring in the direction-tree is discovered and modeled using SVM. Finally, the watermark embedding rules are designed. The original image is not required to take part in extraction, whose algorithm belongs to blind watermark.In this dissertation, the proposed watermarking algorithm are carried out a lot of invisibility and robustness of the simulation experiment. The results show that two kinds of image watermarking algorithms do not all have good visibility of the common image processing operations are highly robust.