Research Of Digital Watermarking Algorithm Based On Non-Uniform B-Spline Wavelets

The revolution of digital information has brought about many deeply changes in past decades, and caused many new challenges and opportunities to our society and our lives. For example, how to protect copyright of digital information has absorbed much attention. Digital watermarking may be a good way to resolve the issue, and thus has become a hot research topic.The applications of non-uniform B-spline wavelets in geometric modelling can be found in many literatures, but that in digital watermarking is still blank. In this paper, both the fundamental theories of non-uniform B-spline wavelets and the method of constructing non-uniform B-spline wavelets with minimal support are introduced in detail.Then it is applied to digital watermark technology. The main works in this article are as follows:In chapter 3, the decomposition and reconstruction of one-dimensional non-uniform B-spline wavelets is introduced and then is extended to two-dimensional images. A method for determining knots in decomposition and reconstruction for images is proposed, the formulas for decomposition and reconstruction of image are given. Its application to resizing image is demonstrated by examples.In chapter 4, taking advantage of the properties of non-uniform B-spline wavelets and considering the balance between invisibility and robustness, two new algorithms for digital image watermarking based on non-uniform B-spline wavelets are proposed. The one merit of the two algorithms is that not only 1/2k size of original image, but also arbitrary binary watermark image which is smaller than carrier one can be embedded, and can be extracted blindly. The other merit is the algorithms are more robust against cropping than that based on haar wavelets.In chapter 5, a new audio watermarking algorithm based on non-uniform B-spline wavelets transform is presented. The experiments show that the algorithm assures the watermark's imperceptibility and is robust against common audio process such as low-pass filtering, adding noise, resampling, requantization, cropping, etc.