本书通过为读者提供有关多媒体司法鉴定年轻领域的发展的更广阔视野，描述了多媒体指纹识别的框架。EURASIP Book Series on Signal Processing and Communications, Volume 4Multimedia Fingerprinting Forensics forTraitor TracingK. Ray liu, Wade trappe, Z. Jane Wang, Min Wu, and Hong ZhaoHindawi Publishing Corporationhttp://www.hindawi.comEURASIP Book Series on Signal Processing and CommunicationsEditor-in-Chief: K.J. Ray liuEditorial Board: Zhi Ding, Moncef Gabbouj, Peter Grant, Ferran Marques, Marc Moonen,Hideaki Sakai, Giovanni Sicuranza, Bob Stewart, and Sergios TheodoridisHindawi Publishing Corporation410 Park Avenue, 15th Floor,#287 pmb, New York, NY 10022, USANasr City Free Zone, Cairo 11816, EgyptFax:+1-866-HINDAWI (USA Toll-Free)c 2005 Hindawi Publishing CorporationAll rights reserved. No part of the material protected by this copyright notice may be reproduced orutilized in any form or by any means, electronic or mechanical, including photocopying, recording,or any information storage and retrieval system, without written permission from the publisher.ISBN977-5945-18-6DedicationTo Our familiesContentsPreface1. Introduction2. Preliminaries on data embedding2. 1. Content protection via digital watermarking2.1.1. Major applications and design requirements2. 2. Basic embeddingapproaches7892.2. Robust additive spread-spectrum embedding2. 2.1. Overview of spread-spectrum embedding122.2.2. Distortion and attacks against robust embedding132.2 3. Mathematical formulation2.2. 4. Alternative detection statistics172.2.5. Exploiting human visual properties212.3. Employing spread-spectrum embedding in fingerprinting233. Collusion attacks253. 1. Introduction to collusion attacks263. 1. 1. Linear collusion attacks263. 1.2. Nonlinear collusion attacks3. 2. Introduction to order statistics293.2.1. Distribution of order statistics303.2.2. Joint distribution of two different order statistics303.2.3. Joint distribution of order statistics andthe unordered random variables313.3. Multimedia fingerprinting system model333.3.1. Fingerprinting systems and collusion attacks333. 3.2. Performance criteria353. 4. Statistical analysis of collusion attacks363.4.1. Analysis of collusion attacks363.4.2. Analysis of detection statistics413.4.3. System performance analysis423.5. Collusion attacks on Gaussian-based fingerprints3.5.1. Unbounded Gaussian fingerprints433.5.2. Bounded Gaussian-like fingerprints483.6. Preprocessing of the extracted fingerprints523.7. Experiments with images573.8. Chapter summary61Contents4. Orthogonal fingerprinting and collusion resistance4.1. Collusion resistance analysis654.1.1. The maximum detector64.1.2. The thresholding detector4.2. Extensions to other performance criteria4.3. Extensions to other types of attacks834.4. A practical estimator for the amount of colluders884.5. Experiments with images904.6. Efficient fingerprint detection using tree structure4.6.1. Tree-structured detection strategy944.6.2. Experiments on tree-based detector4.7. Chapter summary95. Group-oriented fingerprinting1015.1. Motivation for group-based fingerprinting1025. 2. Two-tier group-oriented fingerprinting system1055.2. 1. Fingerprint design scheme1055.2. 2. Detection scheme1065.2.3. Performance analysis5.3. Tree-structure-based fingerprinting system1215.3. 1. Fingerprint design scheme1215.3. 2. Detection scheme1225.3.3. Parameter settings and performance analysis1245.4. Experimental results on images1325.5. Chapter summary1356. Anticollusion-coded(ACC) fingerprinting1376.1. Prior work on collusion-resistant fingerprinting for generic data 1396.2. Code modulation with spread-spectrum embedding1426.3. Combinatorial designs1436. 4. Combinatorial-design-based anticollusion codes1486.4.1. Formulation and construction of Acc codes1496.4.2. Examples of BIBD-based ACC1506.4.3. ACC coding efficiency and Bibd design methods1526.5. Detection strategies and performance tradeoffs1546. 5. 1. Hard detection1566.5. 2. Adaptive sorting approach1576.5.3. Sequential algorithm1576.6. Experimental results for ACC fingerprinting1586.6. 1. ACC simulations with Gaussian signals1586.6.2. ACC experiments with images1636.7. A unified formulation on fingerprinting strategies1646.8. Chapter summary1687. Secure fingerprint multicast for video streaming1717.1. Secure video streaming172Contents7. 2. Prior art in secure fingerprint multicast1737.3. General fingerprint multicast distribution scheme1747. 4. Joint fingerprint design and distribution scheme1767.4.1. Comparison of fingerprint modulation schemes1777.4.2. Joint fingerprint design and distribution1807.4.3. Addressing the computation constraints1857.5. Analysis of bandwidth efficiency1867.5.1."Multicast only scenario1867.5.2. General fingerprint multicast scheme1877.5.3. Joint fingerprint design and distribution scheme1917.6. Robustness of the embedded fingerprints1947.6.1. Digital fingerprinting system model1947.6.2. Performance criteria1957.6.3. Comparison of collusion resistance1957.7. Fingerprint drift compensation1997.8. Chapter summary2028. Fingerprinting curves2058.1. Introduction2058. 2. Basic embedding and detection2088.2.1. Feature extraction2088.2.2. Fingerprinting in the control-point domain2108.2.3. Fidelity and robustness considerations2128.2.4. Experiments with simpldecurves2158.3. Iterative alignment-minimization algorithm forrobust fingerprint detection2198. 1. Problem formulation2218.3.2. Iterative alignment-minimization algorithm2228.3.3. Detection example and discussion2258.4. Experiments with maps2288.5. Chapter summary237Bibliogr239Index251PrefaceMultimedia is becoming an integral part of our daily life. It is a means for us tocommunicate important information with each other, as well as a way to expressour creative sides. The information and art contained inside media have economicvalue, personal value, and often broader impacts on the general welfare of oursociety Consequently, multimedia is a form of digital information that must beprotectedThis book is about protecting the economic and sensitive nature of multimedia. Since the Internet has become increasingly widespread and now reaches intoour everyday actions, it is easy to foresee that our modern communication networks will become the means for distributing multimedia content This distribution will take many forms, ranging from a deceptively simple download-and-playmodel where a single consumer is the end-target for that content to streamingmodes of operation where content is being enjoyed simultaneously by many con-sumers. Regardless of how you look at it, the future of multimedia is closely tied tothe pervasiveness of our communication infrastructure. It therefore seems naturalto protect multimedia by securing its distribution across these networks, that is,y employing the methods of network securityAlthough securing the network and protecting the data crossing the networfrom eavesdropping is certainly essential for protecting multimedia, it is nonetheless a generic problem with generic solutions. Network security methods are im-portant to many other applications, such as electronic commerce and computer se-curity, in addition to being important to multimedia security. However, this book,Multimedia Fingerprinting Forensics for Traitor Tracing, is not about securing thecommunication infrastructure that will deliver multimediaRather, this book focuses on the issue of protecting multimedia content whenit is outside the realm of cryptography and network security. It is now relativelyeasy for adversaries to access multimedia content after it has been decrypted. Adversaries may now alter and repackage digital content. Therefore, ensuring thatmedia content is employed by authorized users for its intended purpose, regard-ess of how it was delivered, is becoming an issue of eminent importance for bothgovernmental security and commercial applications. As such, this book is aboutissues that are unique to multimedia and focuses specifically on how multimediaunlike generic data types, can be protected by using fingerprint signals that are invisibly embedded inside the multimedia to trace and deter unauthorized contentredistribution That is, this book is about the rather nascent field of multimediaforensics, where the goal is to track and identify entities involved in the illegalmanipulation and unauthorized usage of multimedia content ultimately, a solidfoundation for media forensics will deter content fraudPrefaceThis book is targeted at an audience that is familiar with the fundamentalsof multimedia signal processing and will teach the reader about the tools neededto build, analyze, and deploy solutions that will protect a variety of multimediatypes. It, therefore provides foundational material intended to assist the digitalrights management(DRM)engineer understand technologies that complementtraditional cryptographic security methodIn this book, we will review a few major design methodologies for collusionresistant fingerprinting of multimedia and highlight common and unique issues ofvarious different fingerprinting techniques. The goal is to provide a broad overviewof the recent advances in fingerprinting for tracing and identifying colluders. Wewill first provide background on robust data embedding, upon which multimedia fingerprinting system is built. We will then introduce the basic concepts offingerprinting and collusion and provide a discussion on the various goals associated with fingerprint design and colluder tracing. Detailed discussions are thenprovided on two major classes of fingerprinting strategies, namely, orthogonal fingerprinting and correlated fingerprinting, where the latter involves the design ofsuitable codes that are employed with code modulation to create the fingerprintsAs part of our discussion, we will arrive at a unified view of fingerprint design thatcovers orthogonal fingerprints, coded fingerprints, and other correlated fingerprints. After concluding the discussion of fingerprint design methodologies,wewill explore two applications of fingerprinting. We will explore the migration ofmultimedia forensic technologies to networks, whereby the fingerprinting processwill be integrated in core multicast functionality to provide drm solution suitablefor streaming delivery of content. Next, we will examine the protection of a type ofmultimedia content that has, until recently, been left unprotected by multimediasecurity solutions. In particular, we will explore the design of fingerprints for digital curves and maps and exploit the unique properties of digital curves in order todevise fingerprinting solutionsWe would like to thank Ms. Hongmei Gou, a Ph D student in the Univer-sity of Maryland, for her contribution and involvement in preparing the draft ofChapter 8. The results presented in this book have been, in part, supported by theational science foundation and the air force research laboratories We wouldlike to thank these organizations for the support to explore and develop this exciting research areaK.. Ray liuWade TrappeZ. Jane WangHong zhao