【作者】 曹薇；

【导师】 高隽；

【作者基本信息】 合肥工业大学 ， 信号与信息处理， 2004， 硕士

【摘要】 数字图像水印技术是信息隐藏的一个分支，它利用人眼视觉冗余，将标志信息作为水印嵌入到被保护图像之中，是数字图像版权保护的一个新兴方法，也是数字产品版权保护领域中的一个研究热点。利用数字图像水印技术嵌入的被隐藏信息具有不可见性、鲁棒性和安全性等特点，可以不破坏被保护图像的图像质量，同时使得被保护图像在传播、复制和非法篡改之后仍然能够被判断出其版权归属。而随着Internet和电子商务的普及，用于数字产品版权保护的数字图像水印技术作为传统加密技术的有效补充，正越来越受到各方的重视，并逐渐发挥出越来越多的作用。 本文开展“数字图像水印技术及其算法的研究”主要工作如下： 1．收集和整理了大量资料，总结了近年来国内外在数字水印领域内的研究成果和进展，对水印技术的各种算法进行了较为系统的研究。 2．针对现有数字图像水印技术中存在的用途不明确、水印形式抽象以及水印嵌入区间选取较为复杂的问题，提出一种专门用于版权保护的数字图像水印算法。在此算法中，以二值图像作为水印，具有较强的直观性；利用经典的数字图像处理技术中的离散余弦变换和边缘检测算法确定水印嵌入区域并完成水印嵌入。实验表明，此算法可以保证水印不可见性和鲁棒性要求，适用于数字图像的版权保护。并且，我们将算法进一步细化实现了一个较为完整的数字图像水印平台，该平台专用于数字图像的版权保护领域，具有良好的操作界面和实用性。此技术申报了发明专利，并已通过初审。 3．针对原始图像在水印检测时难以获得的实际问题，以及解决这类问题的现有盲水印算法大多存在水印形式直观性差、检测时阈值难以确定的缺点，提出一种基于盲信号分离的盲水印算法。该算法将盲信号分离问题和离散小波变换相结合用于数字图像水印技术：水印嵌入时，以二值图像作为水印，嵌入到被保护图像的小波域中，保证了水印的鲁棒性和不可见性；水印提取时利用盲信号分离算法在没有原始图像的情况下将具有一定分辨率的水印图像提取出来，直观方便。实验证明，这种基于盲信号分离的盲水印算法效果令人满意。更多还原

【Abstract】 Digital image watermark, as a branch of data hiding technology, imbeds symbol information into the protected image using the redundancy attribute of Human Visual System. As a new methodology of copyright protection of digital image, it has become a hotspot in that field recently. Because of the attributes of invisibility, robustness and safety for the imbedded information, it is possible that the quality of the imbedded image is not destroyed after imbedding operation, meanwhile the ascription of copyright of the protected image is still available after some processing. Furthermore, with the development of the Internet and electronic business, digital image watermark, as an effective complementarity for the traditional cryptography technology, is getting more and more interest now.Our work in this paper is mainly about:1. A survey of the research and the progress in the domestic and aboard field of digital image watermark and a systematic research on the various algorithm of digital watermark have been made on the basis of the large number of collective literature.2. A novel algorithm of digital image watermark for the copyright protection is proposed to solve the existing problem in current technology of digital image watermark such as uncertain purpose, abstract form of watermark and complexity in the selection of imbedded zone. In this algorithm, the intuitionistic property is assured by using the binary image as the watermark; Discrete Cosine Transform and classical edge detection algorithm are adopted to determine the imbedded zones for the watermark. It is experimentally proved that invisibility and robustness of watermark are assured by our algorithm and present algorithm fits for the copyright of digital image. Moreover further work has been made to implement our algorithm by building a complete platform of digital image watermark which is special for the copyright of digital image with its friendly interface and high value in practice. This technology has been applied for the patent, and now the initial trial is passed.3. A novel blind watermarking algorithm based on the Blind Signal Separation(BSS) is activated from the fact that the source image is often not available in practice and existing algorithm for this problem is not powerful because of the abstract form of watermark and difficulty in determining the threshold. In this algorithm, BSS and Discrete Wavelet Transform are combined: Binary image is used as watermark and is embedded into the wavelet domain to ensure the robustness and invisibility of watermark; BSS is applied in the extraction of watermark to get the watermark with definite resolution on the condition of lack of source image. It is experimently proved that our blind watermark algorithm based on BSS is satisfying.更多还原