【作者】 黄豫蕾；

【导师】 陈树中；

【作者基本信息】 华东师范大学 ， 系统分析与集成， 2004， 硕士

【摘要】 随着Internet的迅速发展，各种各样的数字多媒体信息包括文本、图像、音频、视频等通过网络广泛传播。但是网上的信息可以被方便地复制和修改，因特网上的侵权问题变得越来越严重。因此，如何保护作品的版权已经受到人们的广泛重视。为了解决这个问题，数字水印技术应运而生。 数字水印技术将特定信息嵌入数字产品中，是一种新型的版权保护机制。借助数字水印可以证明原创作者对其数字产品的所有权，并作为鉴定、起诉非法侵权的证据。一般而言，数字水印算法按嵌入域可分为空间域方法和变换域方法。空间域方法比较简单，但这类方法对于极小的改动都很脆弱。在许多情况下，即使由于有损压缩的很小变化也能使整个隐藏信息丢失。变换域方法是在宿主图像的重要区域嵌入水印，因而比空间域方法能够更好地抵抗压缩、剪切和其它一些图像处理操作的攻击。事实上，这类算法不仅能更好地抵抗各种信号处理操作，而且还保持对人类感观系统的不可感知性。目前主要有两种变换域方法，一种是使用离散余弦变换(DCT)，另一种是使用离散小波变换(DWT)。 本文主要讨论小波域的数字水印算法。小波分析是目前国际上公认的最新时间-频率分析工具。在小波系数中嵌入水印，可保证水印的鲁棒性较好。本文提出了一种DWT域的非盲数字水印算法，该方法根据人眼视觉特性和图像本身的特征决定水印嵌入的位置。实验结果表明水印是不可感知的，并且对于各种图像处理操作是鲁棒的。然后本文根据图像小波系数之间的树结构关系和图像小波系数的统计特征对该算法进行改进，并且将水印作为一幅二值图像来处理，根据JND阈值将小波系数分为两类，对不同的类采用不同的嵌入对策。改进后的算法在水印检测时不需要使用宿主图像。最后本文介绍了一种感兴趣区域的数字水印算法，该方法使特定区域得到更好的保护。更多还原

【Abstract】 With the rapid development of the Internet, various kinds of digital multimedia information including text, image, audio and video etc. have widely distributed through network. But the information on Internet can be easily duplicated and modified. So the copyright infringing problem on Internet is becoming more and more serious. How to provide copyright protection has drawn extensive attention in recent years. To solve this problem, the digital watermarking technique has been investigated.Digital watermarking technique, which embeds special information into digital products, is a new method for copyright protection. With the help of digital watermark, we can prove the originator’s owner to digital products that he creates, and the watermark can be the proof of authentication and prosecuting illegal copyright infringing.Basically, there are two common methods of digital watermarking technique: spatial domain methods and transform domain methods. Spatial domain techniques are easy ways to embed information, but they are highly vulnerable to even small cover modifications. In many cases even the small changes resulting out of lossy compression systems yield to total information loss.Transform domain methods hide watermark in significant areas of the host image that makes them more robust to attacks, such as compression, cropping, and some other image processing, than the spatial methods. However, while they are more robust to various kinds of signal processing, they remain imperceptible to the human sensory system. There are two main transform domain methods, one is to use the discrete cosine transformation (DCT) ; the other would be the use of discrete wavelet transforms (DWT).In this paper, we main discuss the digital watermarking algorithms in wavelet domain. Wavelet analysis is internationally recognized up to the minute tools for analyzing the time-frequency. It has higher robustness to embed digital watermark in wavelet coefficients. A non-blind digital watermarking algorithm is proposed in this paper. The positions of embedding watermark are decided according to human visual system and the image itself. Experimental results show that the watermark is perceptually invisible and the proposed scheme is robust to various image processing operations. Then the method is improved based on the relationship of image wavelet coefficients and statistical characteristic of theimage. The method processes the watermark as a binary image and classifies the wavelet coefficients into two groups using JND (Just Noticeable Distortion) . The watermark is embedded in different way respective to the class. The host image is not required for watermark detection. In the end, the paper introduced a digital watermarking method in regions of interest (ROI) , so special region does receive better protection.更多还原