节点文献

音频数字水印技术的研究

The Research on Digital Audio Watermarking

【作者】 崔丽莉

【导师】 王树勋;

【作者基本信息】 吉林大学 , 通信与信息系统, 2004, 硕士

【摘要】 当前的一场信息革命正席卷全球,网络化浪潮汹涌而来,势不可挡,特别是互联网的爆炸性发展正改变着经济、社会、文化的结构和运行方式,改变着人的思维方式,其广度和深度都是以往任何一次产业革命所无法比拟的。在综合国力竞争十分激烈、国际局势瞬息万变的形势下,一个国家支配信息资源能力越强,就越有战略主动权,而一旦丧失了对信息的控制权和保护权,就很难把握自己的命运,就没有国家主权可言。由于近几年来数字产品知识产权的问题越发突出,数字水印技术也随之倍受科研工作者的关注。数字水印技术的应用非常广泛,可以应用在数字产品的知识产权保护,追踪非法用户,防止篡改原始产品等方方面面。可以预见随着互联网络越来越盛行,数字信息将是未来人们的主要信息交流形式,数字产品给人们带来方便的同时也存在很多不得不解决的问题,如非法盗版以及非法篡改等问题。目前很多学者认为这些问题将随着数字水印技术的完善而得到解决,由此可见数字水印技术在将来信息社会的重要性。而且,当前我国社会信息化正以一日千里的速度前进,对网络与信息安全的需求日益增大。与其它领域不同,网络与信息安全问题必须依靠我国自己的力量来解决。引进国外产品或照搬国外先进技术来解决信息安全问题无异于引狼入室。为此,国家已经明确规定:“信息安全产品一定要立足国内,自主开发”。所以我国学者有责任对数字水印这门学科进行深入的研究,建立自己的一套理论体系,并开发自己的数字水印产品。传统的信息保护手段(例如,密码学等)已经不能满足目前数字化媒体的安全保护。因此,以数字化的信息隐藏理论为核心的技术应运而生。这就是本文的研究背景。本文的主要贡献在于:一 . 搜集整理了大量关于音频数字水印的文献、资料;瑞. 在此基础上,仿真了许多前人的音频水印算法,并将这些算法进行 整理归类,在本文中将这些算法分为了两类:时间域水印算法和变 换域水印算法,同时分析了这两类水印算法的优缺点。江 . 将语音分析中用到的倒谱技术应用到音频数字水印中,并在前人工 97<WP=104>摘 要作的基础上,提出了基于倒谱域的三种数字水印算法:基于信号能量特征算法;基于统计均值调制算法;基于混沌差分 DCSK 调制算法。这三种算法分别针对目前三种不同水印(混沌序列;二值图像;文本信息)的应用而提出,具有很好的不可感知性和鲁棒性。水印的检测过程均为盲检测,符合实际应用的需要,因此具有很好的实用价值。四. 研究了基小波域音频数字水印算法。提出了基于 Sathidevi 的音频压缩方法的水印算法,该算法根据载体信号自身特性,自适应的给出判断阈值,对于抵抗音频压缩攻击(如 MP3)具有很好的鲁棒性。归纳总结了国际音频编码标准 ISO 11172-3 (MPEG-1)标准,借鉴了其中的音频感知模型,提出了基于小波域的水印算法。该算法同样是一个盲检测算法,与已有的时域算法相比,该算法具有较好的鲁棒性和检测精度的准确性。本文的主要工作及其内容安排归纳如下:本论文共分六章。在第一章中,论述了数字水印的研究意义,说明数字水印是信息隐藏这门学科的一个分支。具体介绍了数字水印概念,历史起源,发展现状以及数字水印的特性及其主要应用。第二章是基础知识部分。介绍数字音频水印的基础知识。给出数字水印的嵌入模型和检测模型,音频水印的评估标准。接下来介绍了论文中引用的倒谱的概念,性质及应用。小波的概念,多分辨分析及 Mallat 算法。第三章,介绍经典的数字音频水印算法,将算法分类为:时域和变换域两大类。仿真了其中的最低比特位(LSB)算法,相位编码算法,回声隐藏算法,离散余弦变换(DCT)算法。最后指出目前音频水印算法中存在的问题。第四章,将语音分析中的倒谱技术应用到音频数字水印中,根据嵌入的水印信号的特性不同,提出三种盲检测水印算法,分别为:基于信号能量特征算法,基于统计均值算法,以及基于混沌扩频调制的水印算法。第五章,介绍两种基于小波变换的音频数字水印算法。①提出了基于多分辨分析,及音频压缩的理论的音频水印算法。判断阈值是通过载体信号的特征自适应给出,该算法对于抵抗 MP3 压缩具有很好的鲁棒性。 - 98 -<WP=105>摘 要②介绍了一种音频感知模型,并通过该模型得到水印系统密钥,在此基础上,提出了小波域水印算法。实验结果表明,该系统较好的满足了不可感知性和鲁棒性的要求,与已有的时域算法比较,具有较好的鲁棒性和检测精度的准确性。第六章,对全文进行了总结,同时指出研究中尚有待解决的若干问题,指出下一步将研究的方向,。

【Abstract】 Currently the revolution of information is sweeping all over the world, thetide of network is becoming inundanta, and it is changing our custom ofeconomic, society, culture, people’s thought and so on. The characteristic iswider and deeper than any other industry revolution. The competition amongstates is fury, international situation is fluky, the more information a countrycontrol, the better situation for that country. Recently, the problem of Intellectual Property Rights become more andmore important, and watermarking technology get more and more attention ofscholar. The application of Digital Watermarking technology is wide, which canbe applied in digital product copyright, tracing illegal user, preventing frommodify original product. With the rapid development of Internet, we can forecastthat the digital information will be the main communication in people. Digitalproduct bring a ot of conveniency for people, but at the same time a lot of lproblem that must be solved. For example, the problem of illegal copying andillegal modifying. By now, many scholars think that those problem can be solvedwith the mature of the Digital Watermarking. For this reason, we can see theimportance of the Digital Watermarking in the future. Furthermore, the speed ofour country information society is fast, need more and more security of networkand information. The information security domain is different from otherdomain , which must be done by our own strength. It is dangerous thatintroducing foreign product or advanced technology. For this reason, our countryregulate definitely: The product of information security must be based ourselvesupon. So, the scholars of China have the duty for studying Digital Watermarking,establishing own theory and exploring own Digital Watermarking product. Thetradition means of protection can’t satisfy the requirement for security of digitalmedia, such as cryptography. Accordingly, the technique based on informationhiding emerges as the times require. The above is the research background forthis paper.The creativity task of this paper focuses on the four aspects.1. Collect and clean up a lot of literature about Aduio Digital Watermarking. - 100-<WP=107>Abstract2. On this base, simulate a lot of algorithm of old man, and these algorithm are sorted by us. In this paper, two class audio watermarking algorithms are sorted: time domain watermarking algorithm and transformation domain watermarking algorithm. At the same time, the virtue and defect of these algorithms are analysed.3. The mature cepstrum technology that used in speech analysis is applied in audio watermarking. Based on the ole man research, proposed three kinds of algorithms, they are: based on singal energy character algorithm; based on statistical mean modulation algorithm; Based on DCSK (Differential Chaos Shift Key) algorithm. The three algorithms proposed aims at different watermarks (Chaos, Binary image, text information) application separately, with good imperceptible and robust. The watermarks detection are blind without original signal, answer for the need of practice, so the proposed algorithms have good practicality value.4. Research the audio watermarking in wavelet domain. In the paper proposed a audio watermarking algorithm based on Sathidevi compression. The algorithm based on carry signal character, derive the judge threshold, so have good robust character with the attack of MP3.Sum up the international audio coding and decoding standard ISO 11172-3 (MPEG-1), summarize the audio perceptive model, propose a blind detection algorithm, compare with the existing algorithm in time domain, the proposed scheme can detect watermark accurately. The paper is divided into six chapters: In Chapter one, the study meaning of Digital Watermarking is described. Itexplain that Digital Watermarking is part of

  • 【网络出版投稿人】 吉林大学
  • 【网络出版年期】2004年 04期
  • 【分类号】TP391.4
  • 【被引频次】7
  • 【下载频次】589
节点文献中: 

本文链接的文献网络图示:

本文的引文网络