节点文献
抗时间同步攻击的数字视频水印方法研究
【作者】 孙建德；
【导师】 刘琚；
【作者基本信息】 山东大学 ， 通信与信息系统， 2005， 博士
【摘要】 近年来,随着网络和多媒体技术的迅猛发展,数字多媒体产品的生产、复制、获得和传播变得越来越容易。然而,这在方便了合法使用者的同时,也为出于不同目的的盗版或侵犯生产者和消费者合法权益的非法存取、复制、篡改和传播等侵权行为提供了方便,这就使得数字产品的版权问题成为了亟待解决的问题。数字水印技术就在这种背景下应运而生,并成为当前进行数字产品版权保护的主要方法。数字水印技术就是通过一定的算法将版权信息嵌入到被保护的数字载体信号(如图像、音频、视频等)中来证明版权归属或侵权行为的技术,它已经成为多媒体信号处理领域中的一个研究热点。数字水印技术的研究和应用对版权保护、数据认证和完整性检验、使用控制、广播监控、用户跟踪等方面都有非常重要的理论意义和应用价值。 独立分量分析(ICA—Independent Component Analysis)是在盲信号处理的研究过程中出现的一种信号处理和数据分析的方法。利用它,可以在不知道源信号和传输信道参数的情况下,根据输入源信号的统计特性,仅通过观测信号恢复或提取源信号。它是当今信号处理领域中的一个研究热点,在语音信号处理、阵列信号处理、生物医学信号处理、移动通信、图像处理和噪声消除等方面有着广泛的应用。 本文总结了前人对数字水印和ICA的研究,基于ICA在信号分离、特征提取方面和数字水印问题的相通性,以ICA为主要数学工具,对图像/视频水印问题进行了系统的研究,在获得一些基于ICA的图像/视频特征后,主要针对数字水印对于同步性攻击的鲁棒性问题提出了一系列基于ICA的图像/视频水印方案。主要成果概括如下: 1、提出直接进行水印叠加和提取的图像水印方案。该方案根据ICA模型中信号的混叠/分离与数字水印嵌入/提取过程之间的相通性,以线性叠加的方式将水印嵌入到图像的时(空)域上,检测的时候通过ICA的方法来提取水印。虽然该方案是非盲检测的,但是它的嵌入和提取过程简单、稳定,嵌入的水印信息量大,水印的不可见性好,而且在密钥图像的辅助下,对于各种常见攻击都表现出较好的鲁棒性;
【Abstract】 With the development of network and multimedia techniques, digital multimedia products can be produced, replicated, accessed and distributed more and more easily. It is convenient for the legal customers, whereas it provides the chance to pirates to infringe the copyright for various illegal purposes. That makes copyright protection an urgent issue to be resolved. The technique of digital watermark is being developed under this circumstance, and has been the main method of multimedia copyright protection. Digital watermarking is to embed the copyright information, according to some particular algorithm, into digital host signal, such as image, audio, video, and so forth, to authenticate the copyright. And it has been the focus in the field of multimedia signal processing. The research and application of digital watermark will be of great value both in theory and in application for copyright protection, integrity authentication, access control, broadcast monitoring, customers tracking, and the like.Independent Component Analysis (ICA) is a method of signal processing and data analysis developed in the research of blind signal processing. By the approach of ICA, even without any information of the source signals and the coefficients of transmission channel, people can recover or extract the source signals only from the observations according to the stochastic properties of the input signals. It is a present active research of signal processing. It has lots of applications in speech signal processing, array signal processing, biological and medical signal processing, mobile communication, and image processing, to name just a few.On the basis of previous research on digital watermarking and ICA, this dissertation analyzes the similarity between the ICA model and the watermarking system, and the feasibility of applying feature extraction by ICA on digital watermarking, based on which, ICA is used as the tool to get some kinds of features of image/video. After analyzing the image/video watermarking systemically, the dissertation presents several image/video watermarking schemes based on ICA, mainly to resist the synchronization attacks. The dissertation is composed of the following parts:1. An image watermarking scheme in spatial domain is proposed in chapter 4. Founded on the similarity between the ICA model and the watermarking system, in this scheme, the watermark is embedded into an image by being linearly added into the spatial domain directly, and is extracted by ICA during watermark detection. Though the extraction of this scheme is non-blind, the embedding and extraction can be performed simply and stably. And in this scheme, the watermark has the same amount of information as the host image, and its imperceptibility is good. Furthermore, thanks to the key image, the watermark is robust against most of the common attacks on image watermark.2. An image watermarking scheme based on image independent block feature (IIBF) is also described in chapter 4. In this scheme, the image is divided into uniform blocks, and the IIBFs can be obtained from these blocks by using ICA. Firstly, the primary scheme is attained, in which the watermark is embedded into the least significant components of IIBF to achieve imperceptibility. Founded on the idea of the second generation watermark, DCT and the idea of adaptation are adopted to make an improvement, in which the watermark is embedded into the low frequency in DCT domain of IIBF. At the same time, the order indetermination of ICA is taken into account and the effect on waterjnark caused by attacks is estimated. The simulations show that the robustness of the improved scheme to cropping, translation, and rotation, to name just a few, is improved greatly.3. In chapter 5, a video watermarking scheme based on 3D video independent block feature is presented. In this scheme, which is looked on as the extension of the IIBF-based image watermarking scheme, the video is divided firstly according to the structure of group of picture (GOP), and each GOP is divided into 3D-blocks, from which video independent block feature (VIBF) can be extracted by using ICA, and then the watermark is embedded into the slices of VIBF. The simulations show that in the case of frame dropping, the video can be resynchronized with the watermark being extracted correctly. Certainly, the scheme is robust against the common video watermark attacks, e.g. MPEG compression, collusion attack. In addition, the size of VIBF and the selection of slices of VIBF are discussed.4. A video watermarking scheme based on the relative motion of consecutive frames is explored in chapter 6. In this scheme, ICA is applied to extract video independent dynamic feature (VIDF) and video independent static feature (VISF) from every two consecutive frames, and watermark is embedded into the VIDF. In order to resist the copy attack and temporal desynchronization, the signature of I-frame of each GOP, which can represent GOP exclusively, is selected as the watermark. The simulation results demonstrate the value of this kind of watermark. The scheme is robust not only to the common video watermark attacks, e.g. MPEG compression, collusion attack, but also to temporal desynchronization, such as frame dropping, frame swapping, etc. In addition, the watermark based on signature of GOP can resist copy attack effectively.5. In chapter 7, a video watermark scheme based on video content is proposed. In this scheme, the video is divided into segments based on different scenes, and principal independent component analysis (PICA) is used to extract the video independent content feature (VICF) from each segment. And the watermark is embedded into these VICFs. In chapter 7, how to perform PICA is introduced, and the property of VICF is analyzed. Different from the conventional methods based on temporal template or particular structure of watermark, the method of resistance to temporal desynchronization is based on video content. This scheme is robust to most of the common video watermarking attacks, e.g. filtering, MPEG compression, noising, as well as to temporal desynchronization, e.g. frame dropping, frame swapping, frame averaging, and transcoding between MPEG2 and MPEG4. In addition, its robustness to the temporal mis-segmentation caused by fade in/out is discussed.In this dissertation, ICA is used as the mathematical tool, to get a series of image/video features. Founded on these features, several image/video schemes are proposed to improve the robustness of watermarking, especially the robustness to synchronization attacks. Among these schemes, the VICF-based video watermarking scheme resists temporal desynchronization from the perspective of video content, which is different from the conventional idea based on template or particular structure of watermark. In conclusion, the application of ICA on watermarking provides novelmethods of watermarking, and novel application area for ICA. Finally, the dissertation summarizes the to-be-resolved problems in watermarking research, and the emphases of further research. Besides, it provides an outlook of the application of watermark to multimedia retrieval.