Multimedia Data Hiding

227 Pages·2003·7.693 MB·English

by  Min Wu, Bede Liu (auth.)

Multimedia Data Hiding
Springer
Science+Business Media,
LLC
Min Wu Bede Liu
Multimedia Data Hiding
With 92 Illustrations
, Springer
MinWu Bede Liu
Department of Electrical Department of Electrical
Engineering Engineering
University of Maryland Princeton University
College Park, MD 20742 Princeton, NJ 08544
USA USA
[email protected] [email protected]
Library of Congress Cataloging-in-Publication Data
Wu, Min, 1974-
Multimedia data hiding / Min Wu, Bede Liu.
p. cm.
Includes bibliographical references and index.
ISBN 978-1-4419-2994-5 ISBN 978-0-387-21754-3 (eBook)
DOI 10.1007/978-0-387-21754-3
1. Multimedia systems. 2. Data encryption (Computer science) 3. Computer
security. 1. Liu, Bede. II. Title.
QA 76.575 .W85 2002
006.7-dc21 2002030240
Printed on acid-free paper.
© 2003 Springer Science+Business Media New York
Originally published by Springer-Verlag New York, Inc in 2003
Softcover reprint of the hardcover 1s t edition 2003
AII rights reserved. This work may not be translated or copied in whole or in part without the
written permission ofthe publisher (Springer Science+Business Media New York), except for
brief excerpts in connection with reviews or scholarly analysis. Use in connection with any
form of information storage and retrieval, electronic adaptation, computer software, or by
similar or dissimilar methodology now known or hereafter developed is forbidden.
The use in this publication oftrade names, trademarks, service marks, and similar terms, even
if they are not identified as such, is not to be taken as an expression of opinion as to whether
or not they are subject to proprietary rights.
987 6 5 4 3 2 1
www.springer-ny.com
To Our Families
Preface
The digital information revolution has brought about profound changes in
our society and our life. New devices and powerful software have made it
possible for consumers worldwide to create, manipulate, share, and enjoy
the multimedia information. Internet and wireless networks offer ubiquitous
channels to deliver and to exchange multimedia information for such pur
poses as remote collaboration, distant learning, and entertainment. With
all these advances in multimedia coding and communication technologies
over the past decade, the major hurdle for allowing much broader access of
multimedia assets and deployment of multimedia services no longer lies with
bandwidth-related issues, but with how to make sure that content is used
for its intended purpose by its intended recipients. The core issue then be
comes the development of secure management of content usage and delivery
across communication networks.
Data hiding and digital watermarking are promising new technologies for
multimedia information protection and rights management. Secondary data
can be embedded imperceptibly in digital multimedia signals for a variety
of applications, including ownership protection, authentication, access con
trol, and annotation. Data hiding can also be used to send side information
in multimedia communication for providing additional functionalities or for
enhancing performance. The extraction of the embedded data mayor may
not need knowledge of the original host media data. In addition to im
perceptibility, robustness against moderate processing such as compression
is also an important consideration. The requirements of imperceptibility,
robustness and the hiding of maximum number of bits are basic for many
data hiding applications. To satisfy these conflicting requirements, attention
viii Preface
must be paid to the visual and hearing perception model and the types of
media data, viz. speech, music, line drawings, signature, natural image, etc.
In addition, different parts of the media data may have significantly differ
ent embedding capacity. How to handle this uneven distribution of capacity
is also a challenge. Another concern of data hiding is the protection against
intentional attacks, attempts aimed at remove or obliterate the hidden data
or watermark.
This book, based on the Ph.D. dissertation of the first author [46], ad
dresses both theoretical and practical aspects of multimedia data hiding,
and tackles both design and attack problems. It is organized in three parts:
Fundamental Issues, Algorithm and System Designs, and Attacks and Coun
termeasures.
In Part I, we identify the key elements of data hiding through a layered
structure. Data hiding is modelled as a communication problem where the
embedded data is the signal to be transmitted. The tradeoff of robustness
versus capacity is studies for two major categories of embedding mecha
nisms. In addition, a comprehensive solution is proposed to address the
problem caused by the unevenly distributed embedding capacity. The ques
tion of constant bit rate versus variable bit rate hiding is also addressed.
In Part II, we present new data hiding algorithms for binary images,
grayscale and color images, and videos. These algorithms can be applied to
a variety of problems, including annotation, tamper detection, copy/access
control, fingerprinting, and ownership protection. The designs presented
here provide concrete examples regarding the choice of embedding mecha
nisms, the selection of modulation/multiplexing techniques for hiding mul
tiple bits, and the handling of uneven embedding capacity. The use of
data hiding in video communication to convey side information for addi
tional functionalities or better performance is demonstrated by the novel
approaches of real-time transcoding and error concealment.
Many data hiding applications operate in a competitive environment
where an adversary has incentives to remove or obliterate the embedded
data. Thus the testing of the robustness and security via attacks is im
portant. In the last part of the book, we discuss a number of attacks and
countermeasures for data hiding systems. The discussion begins with attack
ing three specific types of watermarking schemes, in which full knowledge
of the watermarking algorithms is available. Attention is then turned to
the watermark attack problems for digital music under a unique compet
itive environment, in which the watermarking algorithms are unknown to
attackers. This work is based on our participation in the recent public chal
lenge in the form of attacking four audio watermarking technologies of the
Secure Digital Music Initiative (SDMI).
Acknowledgement Several works included in this book are in collab
orations with our colleagues: Jeffrey Bloom (Sarnoff Corporation), Inge
mar Cox (NEC Research Institute), Scott Craver (Princeton University),
Preface ix
Ching-Yung Lin (IBM T.J.Waston Research Center), Y-M. Lui (Signafy
Inc.), Matt Miller (NEC Research Institute), Peng Yin (Thompson M ultime
dia Laboratory), and Heather Yu (Panasonic Information and Networking
Laboratory). We have also benefited from the discussions and suggestions
from Perry Cook (Princeton University), Persi Diaconis (Stanford Univer
sity), Bradley Dickinson (Princeton University), Edward Felten (Princeton
University), Adam Finkelstein (Princeton University), S-Y. Kung (Prince
ton University), Nasir Memon (Polytechnic University), Shu Shimizu (IBM
Japan), Harold Stone (NEC Research Institute), Edward Tang (Johns Hop
kins University), and Wenjun Zeng (Packet Video Corporation). We are
grateful to the Electronic Frontier Foundation and its legal team for their
effort enabling the inclusion of our work on the SDMI challenge into this
book. Special thanks to Gino J. Scarselli and Grayson Barber.
Peter Ramadge and Sanjeev Kulkarni of Princeton University read through
an earlier version of the manuscript and offered many helpful comments
and suggestions. The first author would also like to thank her colleagues
at the University of Maryland, College Park. Special thanks to K.J. Ray
Liu, Steve Marcus, Andre Tits, and Kawthar Zaki for their support and
encouragement.
We are grateful to the State of New Jersey for R&D Excellence Grant,
to the National Science Foundation for Grant MIP-9408462 and CAREER
Award CCR-0133704, and to Intel Corporation for Technology for Educa
tion 2000 Grant. These grants supported the research work reported in this
book.
We have enjoyed working with the staff at Springer Verlag, New York, es
pecially the Executive Editor on Computing & Information Science, Wayne
Yuhasz, the Associate Editor, Wayne Wheeler, and the Production Editor,
Antonio D. Orrantia.
The first author would like to thank her parents, Xianli Wu and Yiqi Sun,
and her husband, Xiang Yu, for their love, support, and encouragement.
Min Wu and Bede Liu
October 2002
Contents
Preface vii
List of Figures xv
List of Tables xix
1 Introduction 1
1.1 Overview of Multimedia Data Hiding. 2
1.2 Book Organization .......... . 6
1.2.1 Fundamental Issues and Solutions 6
1.2.2 Algorithm and System Designs 7
1.2.3 Attacks and Countermeasures ... 10
I Fundamental Issues 13
2 Preliminaries 15
2.1 Data Hiding Framework ...... . 15
2.2 Key Elements and A Layered View . 16
3 Basic Embedding Mechanisms 19
3.1 Two Basic Embedding Mechanisms. 20
3.1.1 Probability of Detection Errors 23
3.2 Embedding Capacity . . . . . . . . . . 26
3.2.1 Capacity for Type-I Embedding. 27
xii Contents
3.2.2 Capacity of Type-II Embedding. . . . . . . 27
3.2.3 Capacity Comparison for Type-I & Type-II 29
3.2.4 Extensions and Discussions . . . . . . . . 31
3.3 Techniques for Embedding Multiple Bits. . . . . 33
3.3.1 Modulation and Multiplexing Techniques 33
3.3.2 Comparison................. 35
3.4 Chapter Summary . . . . . . . . . . . . . . . . . 37
3.5 Appendix - Derivations of Type-II Embedding Capacity 38
4 Handling Uneven Embedding Capacity 41
4.1 Quantitative Model for Uneven Embedding Capacity. 42
4.2 Constant Embedding Rate (CER) ........... 44
4.2.1 Backup Embedding ............... 45
4.2.2 Equalizing Embedding Capacity Via Shuffiing . 46
4.2.3 Practical Considerations . 50
4.2.4 Discussions . . . . . . . . . . . . . . . . 50
4.3 Variable Embedding Rate (VER) . . . . . . . . 51
4.3.1 Conveying Additional Side Information 52
4.4 Outline of Examples . . . . . . . . . . . . . . . 53
4.5 Chapter Summary . . . . . . . . . . . . . . . . 53
4.6 Appendix - Generating Shuffiing Table From A Key 54
4.7 Appendix - Analysis of Shuffiing .... 55
4.7.1 Joint Probability of Histogram 55
4.7.2 Mean and Variance of Each Bin. 56
4.7.3 More About E[!ff] . . . . . . . . 58
4.7.4 Approximations for Hypergeometric Distribution 59
4.7.5 More About Var[!ff] ............... 61
II Algorithm and System Designs 63
5 Data Hiding in Binary Image 65
5.1 Proposed Scheme. . . . . . . 67
5.1.1 Flippable Pixels ... 67
5.1.2 Embedding Mechanism 69
5.1.3 Uneven Embedding Capacity and Shuffiing 70
5.2 Applications and Experimental Results. . . . . 74
5.2.1 "Signature in Signature" . . . . . . . . . 74
5.2.2 Invisible Annotation for Line Drawings 74
5.2.3 Tamper Detection for Binary Document 75
5.3 Robustness and Security Considerations . . . . 76
5.3.1 Analysis and Enhancement of Robustness 77
5.3.2 Security Considerations . . . . . . . . . . 81
5.4 Chapter Summary . . . . . . . . . . . . . . . . . 83
5.5 Appendix - Details of Determining Flippability Scores 84