| These years, the multimedia technology and network technology has got rapid development and broad application, so it becomes an impendent problem to be resolved to protect the copyright of the multimedia digital products. Therefore, as a kind of solution, the digital watermarking technology came into being. However, most of the existing watermarking algorithms are proprietary algorithms, and watermark detection or extraction require the original host image to participate in. All these requirements are not available in some practical applications, and it's not real blind extraction. There are more or less problems about the existing watermarking algorithms in terms of invisibility and robustness, therefore, in this paper, I have done some research on the digital image watermarking technology to solve these problems, and it mainly includes the following three aspects:1. Study the classical digital image watermarking algorithm. Simulated and researched the typical digital watermarking algorithms such as Barni algorithm, Kunder algorithm and Xia algorithm, and analysis of advantages and disadvantages of these classic algorithms. They provide a theoretical framework to propose a new algorithm.2. Propose a kind of blind extraction algorithm of image watermarking based on multiple sets of wavelet coefficients. Draws on the concept of multiple sets of wavelet coefficients, based on discrete wavelet transform, and then embed and extract watermark using of multiple relationships between sets of wavelet coefficients. According to the algorithm, we only make small changes to the wavelet coefficients, which greatly improve the invisibility of watermarking system. When extracting the watermark without the original host image, you can achieve true blind extraction.3. Do various common image processing operations and attacks on the watermarked images to test its robustness and optimizate the algorithm to solve these problems that have arisen. |
PDF Full Text Request