【作者】 曹华；

【导师】 周敬利；

【作者基本信息】 华中科技大学 ， 计算机系统结构， 2006， 博士

【摘要】 多媒体信息的数字化极大地提高了信息表达的效率和准确性,网络技术的飞速发展,为传播数字化产品提供了便捷廉价的途径。由于数字信息产品的复制不会产生模拟产品复制时出现的质量降低问题,任意复制传播数字产品的行为变得越来越泛滥。同时,各种图像、音视频编辑软件的应用也变得非常普及,一方面人们可以享受到其带来的便利,但另一方面也使得对数字产品的内容进行篡改变得很容易。为了解决这个问题,数字水印技术通过将一种不可见的标记嵌入到多媒体信息产品中,用以标识原创作者对作品的所有权或对信息的完整性进行验证。从本质上说,数字水印技术是信源编码在信息安全领域的应用,从信息处理的角度看,嵌入载体对象的水印信号可以看作是在强背景下叠加一个弱信号,只要叠加的水印信号强度低于人类视觉或听觉的感知门限,就可以在不引起人类感知的情况下嵌入水印信息。数字图像是网络上最常见到,也是最容易被随意拷贝和修改的多媒体信息,在所有的数字水印技术中,数字图像的水印技术是最受到关注和研究的。目前图像水印的基本实现方法是把载体当做信道,将待隐藏的水印信息用密钥扰乱,用各种方法嵌入到载体中,要求载体的变化不被感知,隐藏的水印信息经过传输和攻击后仍然可以检测提取出来,由此衍生出水印系统中的水印生成、嵌入、提取和攻击四大部分技术,其中水印的嵌入方法对系统性能具有决定性的影响。分形变换利用图像的自相似性进行压缩,可获得较传统算法高得多的压缩比,在未来会得到广泛应用。由于分形变换的实现机制非常特殊,传统的基于DCT、DWT的水印嵌入算法难以在分形变换中实现,因此必须根据分形变换的特点研究有针对性的水印嵌入方法。现有的基于分形变换的水印嵌入技术利用值域子块最佳匹配块的位置不变性,根据最佳匹配块所处的不同寻找区域来嵌入和提取水印,由于有些寻找区域的匹配块并不是最佳的,这些非最佳匹配块参与图像的重构,使水印图像质量有较大的降低,一些改进的算法虽然提高了图像质量,但存在水印容量低、鲁棒性不强的问题。基于对分形变换有关参数的研究,实现一种利用分形变换过程中的对比度因子参数嵌入水印的新方法,可以嵌入较大的水印容量,由于嵌入始终是基于最佳匹配块进行的,重构的水印图像质量很好,同时给出了相应的提取算法,提高了水印提取的准确度。水印检测和提取时需要原始数据和密钥的方案叫做私有水印方案或明水印方案,那些只需要密钥但不需要原始数据的方案叫公有水印或盲水印方案。通常私有水印在经过JPEG压缩和滤波等图像处理后的鲁棒性能比较好,然而私有水印方法在诸如DVD播放的水印探测中是不可行的,因为此时原始数据无法得到,在这种情况下,盲水印方案由于不需要原始数据就可以检测到水印就成为可行的方法,盲水印的缺点是鲁棒性较低,与私有水印相比有较高的虚警率,但它具有的实用性使其依然成为水印的研究热点。对DCT、DWT系数统计模型在水印检测中的性能进行分析,找出对水印系统虚警性能具有决定影响的因素,在此基础上实现两个利用块DCT和DWT变换的盲水印嵌入方法,通过对水印信号进行扩频处理延长水印的长度,在嵌入的过程中引入人眼视觉模型实现自适应嵌入,保证了水印检测过程中的低虚警率。算法简单实用,实验结果表明所提出的两种盲水印算法具有很好的安全性和鲁棒性。以电影、电视剧为代表的影视工业已经成为一个拥有巨大利润空间的产业,它的数字化同样带来了版权保护的问题。视频水印技术作为数字视听产品版权保护系统的核心技术而备受电子厂商和研究者的关注。H.264是一种新的视频编码方法,它具有的很多优良特性使它将成为下一代可应用的视频标准而获得广泛应用,它使用的帧内预测、1/4像素精度运动估计使它可获得比以前视频标准高得多的压缩比和视频质量,但它使用的这些新特性又不利于采用传统视频水印的方法来嵌入水印。在对帧内预测编码模型研究的基础上,实现了一种在H.264编码过程中快速嵌入水印的算法,提取水印时实现了盲检测,并且算法复杂度低可满足视频实时处理的需要。DRM数字版权管理系统是对有形和无形资产版权和版权所有者关系的定义、辨别、交易、保护、监控和跟踪的手段,它为数字媒体的商业运作提供了一套完整的实现手段。数字水印是DRM系统实现版权保护的基本手段,前述的鲁棒图像和视频水印算法可以抵抗试图祛除水印的攻击。协议攻击并不祛除水印,而是通过添加一个伪造的水印来干扰版权的认证,现有的鲁棒水印算法无法抵抗这种攻击,而数字签名是抵抗这种攻击的有效策略,因此将数字水印和PKI结合起来引入到DRM系统中,构造一个具备更完善版权保护功能DRM系统的应用框架,并对其进行了功能描述。更多还原

【Abstract】 The digitization has offered great convenience for the storage and using of the multimedia information and improved the efficiency and accuracy of information expressing greatly at the same time. With the rapidly development of the computer and network technology, Internet has already become the most convenient and the low-priced way to propagate digital products which include electronic publication, digital audio and video, digital cartoon and picture and the limitless business opportunity has appeared with the development of E-commerce technology. But we must also face a new severe challenge which is duplicating, spreading and distorting illegally of digital products because the case of quality degradation which occurred frequently in analog reproduction do not take place while duplicating the digital information products, it is totally lossless to reproduce the digital information. With the performance strengthens and price reducing constantly of the various kinds of digital hardware products, the wanton behavior of duplicating the digital products overflows more and more. This is very unfavorable to those authors of digital works. Meanwhile the performance of the personal computer is stronger and stronger now, it is also popularized very much that the application of various kinds of software for editing picture, audio and video, on one hand people can enjoy the facility brought by these software, on the other hand tampering content of digital products become very easy.In order to solve this problem, digital watermarking technology is used to label the original creator’s ownership or proving to the authenticity and integrality of the digital works through embedding a kind of invisibility mark into the multimedia information products. As a new effective technological means to the copyright protection and content authenticity of digital information products, the digital watermarking technology has already become the research focus of the field of information security nowadays.The digital image is the most common multimedia information on the network, which is also very easy to be copied and revised. The watermarking of digital image is paid the close attention among all digital watermarking technology and the basic principle and frame of the watermarking of digital image have been analyzed. Although some literatures consider watermarking detection should be the focal point that watermarking study at present, this paper still thinks that the embedding technology of watermarking is the determinant factor which deciding the performance of watermarking system and it is still necessary to study new embedding tactics according to the new transformation methods.There are three kinds of image watermarking embedding algorithm proposed in this paper. One is the watermarking embedding method based on the fractal transformation that is completely different from traditional method to realize the image compression, it contributes to resisting these kinds of attacks based on fractal transformation by embedding watermarking according to its characteristic. The other two are blind watermarking methods that realized by using the DCT and DWT coefficiences of the image blocks. It is of important practical value for this method because the primitive information is no needed when watermarking detection.The movie & TV industry represented by film and TV play has already become an industry with enormous profit space and its digitization has brought the question of copyright protection too because digital video watermarking may become key technology of the digital audio-visual products copyright protection system, which concerned by electronic manufacturer and researcher extremely. H.264 is a new kind of video coding standard and a lot of fine characteristics that it has make it become the applicable video standard of next generation and win extensive application, intra-frame prediction, 1/4 pixel precision movement estimate make it can obtain than before video standard high compression ratio and video quality much. But these new characteristics that it uses are unfavorable to the method to adopt traditional video watermarking to embedding the watermarking again. A quick embedding watermarking tactics based on H.264 coding standard was put forward to in this paper which can detect blindly the watermarking message and has low computer complexity.Digital Rights Management (DRM) system are intended to provide a controlled consumption of digital content to protect the intellectual property rights of the authors and the content providers. It specifies the permissions, constraints, obligations, conditions, and agreements with the holders of intellectual property for content usage. Digital watermarking is the basic method by which the DRM realize copyright protection. Robust watermarking methods can hold out the attacks which attempt to get rid of watermarking, but these methods can’t resist to the watermarking protocol attack which disturbing the copyright authentication through add a counterfeit watermarking rather than to get rid of watermarking. Digital signature is the effective tactics of resisting this kind of attack. Through analyzing structure of DRM system and application of watermark technology among them, the application frame of introducing to DRM system of combining digital watermarking and PKI is proposed in this paper, the function of the application frame was described and the performances of resisting to watermarking protocol attack was analyzed.更多还原