【作者】 熊益群；

【导师】 王让定；

【作者基本信息】 宁波大学 ， 通信与信息系统， 2010， 硕士

【摘要】 随着计算机通信技术和互联网的迅速发展,数字多媒体的传播也越来越便捷,这使得数字作品非常容易被获取和篡改,严重危害了作品拥有者的权利,影响了社会上数字作品服务市场的正常秩序,因此,迫切需要解决数字产品的知识产权保护问题。多年来,数字水印技术作为解决数字产品知识产权保护问题最具优势的方法得到了广泛地研究和应用,并且应用到完整性认证等其它领域,应用的载体包括图像,音频,视频,文本等多媒体载体。一直以来音频水印方法都是以单个音频作品作为载体的,然而,音频作品在流通过程中则大多以专辑等集合形式存在,基于现实需求,本文将对以音频作品集合作为载体的水印算法进行探索和研究。本文基于集群式版权保护和认证的音频水印技术研究,工作和特点主要包括以下几个方面:1、系统地介绍了音频数字水印的概念,基本理论和发展现状,并通过对现有的音频水印算法和水印框架的分析,从全新的角度提出了针对音频集合的集群式水印算法,介绍了集群式水印算法的概念,提出其算法模型和基本框架,根据音频集合的特殊性,探索出对于音频集合的攻击方法。2、利用云模型的模糊性和随机性,提出了一种基于云模型的集群式版权保护算法。将音频集合中所有音频作品小波变换后的特征值映射为相当数量的云滴,将这些云滴组合为云图,并将该云图作为零水印,再根据逆向云发生器计算该云图的三个特征值期望Ex,熵En,超熵He,检测水印时利用这些参数的改变率来评估原始水印和提取水印之间的相似性。根据云模型的模糊性和随机性,其中部分云滴受攻击改变后对整个云图的特征值影响很小,该算法鲁棒性强。3、基于提升小波技术提出了一种既能保护音频集合版权又能对集合中单个音频作品分别进行版权保护的新算法,将音频集合中各个音频作品的提升小波系数分别构造零水印,这些零水印分别用来保护相对应的单个音频作品,为了保护音频集合的版权,则将这些零水印通过随机混合矩阵进行混合可得到整个音频集合的集合零水印。4、对ICA概念进行了深入的分析,探索总结出ICA在水印技术中的典型应用方案,并在此基础上提出了基于ICA的集群式版权保护算法,将ICA同时应用到提取音频集合特征值和水印检测过程中。利用ICA提取音频集合特征值时,把原始音频集合中各个音频的小波系数作为观测信号, ICA方法的特点保证这些音频在音频集合中的顺序不受限制,当顺序改变时,仍能提取出一样的独立分量,并且将水印嵌入到其中两个最重要的独立分量中,这不仅保证了源信号的相互独立性,而且构造了检测水印时需要的足够观测信号;另外将ICA应用到水印检测过程中不需要知道水印嵌入位置,水印嵌入强度等因素。5、提出一种基于ICA和小波包的集群式认证技术,将音频集合进行小波包分解,然后将ICA提取小波包系数的各个独立分量,最后根据分析选择一个合适的独立分量进行水印嵌入,该算法不仅可以在音频集合遭受攻击后进行完整性认证,而且可以对被攻击后的待测音频集合进行初步篡改恢复。6、对Shamir提出的(k, n)密钥分散管理方案进行探索和改进,提出一种基于(k, n/k=n)的集群式认证方法,该方法将水印图像的各个像素值分别进行(k, n)分配,并将各个像素值分配后的n个数值纵向排列为一个矩阵,再将这个矩阵的行向量分别嵌入音频集合中对应的单个音频作品,从而实现了将一个水印信息嵌入到整个音频作品中,提取水印时,设置k=n,只要待测音频集合中有音频被篡改攻击了,变化就能反映到提取的水印图像中,保证算法的脆弱性。更多还原

【Abstract】 With the rapid development of computer communication technology and the internet, the spread of digital media has become more and more conveniently, which makes digital media be easy to be copied and pirated. This circumstance damages the rights of works’owners and impacts the normal service market of multimedia, therefore, it is urgent to protect the copyright of the digital media. In recent years, digital watermarking technology which is used for protecting copyright of the digital media effectively has been widely researched and applied. Digital watermarking technology is also applied for integrity Authentication and the other area. The aimed carrier of watermarking technology contains image, audio, video, text and other multimedia works. In generally, the existed audio watermarking technology is just used for a single audio, however, audio works are often published in the form of audio album(an aggregation), in view of practical needs, in this paper, we will probe and research the watermarking technology which is applied for audio aggregation.The watermarking technology in this paper based on audio aggregation contains some algorithms and features as follows:1, Digital audio watermarking technology, basic theory, and development status was systematically introduced, according to analyze existed audio watermarking algorithm and watermarking framework, we presented audio aggregation watermarking technology from a new point, introduced the concept of audio aggregation watermarking and algorithm model. And in terms of audio aggregation’s particularities, we probe some attacks aimed at audio aggregation.2, According to the fuzziness and randomness of cloud model, we proposed an audio aggregation algorithm based on Cloud model. Wavelet coefficients of original audio aggregate are mapped into Cloud drops, Cloud drops is zero-watermark. And then make use of Reverse-Cloud generator to calculate three numerical characteristics of Cloud model, these characteristics are used for assessing similarity of original watermark and existed watermark. After the attacks, a part of cloud drops have been changed, which has little impact on characteristics of the entire cloud model, so utilize Change rate of these characteristics to detect watermark, the robustness is strong. 3, According to lift-wavelet, we propose an algorithm which not only protects copyright of a single audio but also protects copyright of the whole audio aggregation. Using fast lifting wavelet decomposition technique to extract features from every audio work in aggregation, the features are constructed as zero-watermark of every audio, which is used to implement copyright protection of corresponding audio; compose all zero-watermarks of all audio works into a zero-watermark matrix to get aggregate zero-watermark, which is used to implement copyright protection of the audio aggregation.4, Propose an audio watermarking algorithm based on ICA for implementing copyright protection of audio aggregation, the wavelet coefficients of each audio work in the audio aggregation are set as observation data, and independent components of the observation data are extracted using ICA separate matrix, and then two most important components are chosen to embed the common watermark; ICA is used to detect watermark, it is not necessary to know the embedding location and strength of watermark and other factors5, Propose an audio watermarking algorithm based on ICA for implementing integrity authentication, audio aggregation are put into wavelet packets decomposition, coefficients are extracted by ICA technology, and then a right component are chosen to be embedded watermark. The algorithm not only implements integrity authentication for audio aggregation but also recovers attacked audio aggregation.6, Improve the (k, n) method proposed by Shamir firstly, and propose an audio aggregation watermarking algorithm based on (k, n/k=n) for implementing integrity authentication. Every pixel in watermark image are distributed by (k, n) method, and then embedded into every audio work in audio aggregation, so one watermark image are embedded into an audio aggregation, when extracting watermark, set k=n, once there are audio works in aggregation are attacked, the extracted watermark image will change correspondingly, which enhances fragility of the algorithm.更多还原