Summary: | 碩士 === 國立臺灣大學 === 電信工程學研究所 === 89 === The world has evolved into a multimedia age with booming internet connectivity, which enables us to carry and distribute digital data without any difficulty. However, this convenience also facilitates illegal uses of copyrighted data. To prevent pirating and deal with the intellectual property problem, the watermarking techniques are researched and developed to provide the author and owner with another method to declare their ownership.
Two watermarking schemes on color images and digital audio data are proposed in this thesis. For color image watermarking, the watermark is inserted into the index domain of a palette-based image. For audio watermarking, we will embed audio clips into host audio as watermarks other than the commonly used binary sequences in a spread spectrum manner. In both methods, the watermark is embedded repeatedly but the repeating process has different influences on the two forms of media.
This thesis consists of six chapters. A general introduction to digital watermarking is made in chapter 1. In chapter 2, the concept of watermarking will be explained, classifications of watermarks will be given, and related works in digital image watermarking will be summarized. The proposed image watermarking technique on palette-based color images using neural network training will be presented in chapter 3. After that, related works in audio watermarking will be introduced in chapter 4. In chapter 5, the concept of sequences of autocorrelation function with zero sidelobe will be explained, our audio watermarking scheme based on the spread spectrum concept will be presented, and experimental results using a new quantitative approach will be displayed. Finally, a concluding remark and future research directions shall be given in chapter 6.
|