本资源来自于互联网，仅供学习研究之用，不可涉及任何商业用途，请在下载后24小时内删除。
著作权归原作者或出版社所有。未经发贴人conanwj许可，严禁任何人以任何形式转贴本文，违者必究！
Distributed Source Coding: Theory, Algorithms and Applications
Authors
Pier Luigi Dragotti
Publisher: 2009, Elsevier
Pages: 338
ISBN 13: 978-0-12-374485-2
Introduction
In conventional source coding,a single encoder exploits the redundancy of the source
in order to perform compression. Applications such as wireless sensor and camera
networks, however, involve multiple sources often separated in space that need to
be compressed independently. In such applications, it is not usually feasible to first
transport all the data to a central location and compress (or further process) it there.
The resulting source coding problem is often referred to as distributed source coding
(DSC). Its foundations were laid in the 1970s, but it is only in the current decade
that practical techniques have been developed,along with advances in the theoretical
underpinnings.The practical advanceswere,in part,due to the rediscovery of the close
connection between distributed source codes and (standard) error-correction codes
for noisy channels. The latter area underwent a dramatic shift in the 1990s, following
the discovery of turbo and low-density parity-check (LDPC) codes. Both constructions
have been used to obtain good distributed source codes.
In a related effort,ideas from distributed coding have also had considerable impact
on video compression, which is basically a centralized compression problem. In
this scenario, one can consider a compression technique under which each video
frame must be compressed separately, thus mimicking a distributed coding problem.
The resulting algorithms are among the best-performing and have many additional
features,including,for example,a shift of complexity from the encoder to the decoder.
This book summarizes the main contributions of the current decade.The chapters
are subdivided into two parts. The first part is devoted to the theoretical foundations,
and the second part to algorithms and applications.
Chapter 1, by Eswaran and Gastpar, summarizes the state of the art of the theory
of distributed source coding, starting with classical results. It emphasizes an important
distinction between direct source coding and indirect (or noisy) source coding:
In the distributed setting, these two are fundamentally different. This difference is
best appreciated by considering the scaling laws in the number of encoders: In the
indirect case, those scaling laws are dramatically different. Historically, compression
is tightly linked to transforms and thus to transform coding. It is therefore natural
to investigate extensions of the traditional centralized transform coding paradigm to
the distributed case. This is done by Chaisinthop and Dragotti in Chapter 2, which
presents an overview of existing distributed transform coders. Rebollo-Monedero and
Girod, in Chapter 3, address the important question of quantization in a distributed
setting. A new set of tools is necessary to optimize quantizers, and the chapter gives
a partial account of the results available to date. In the standard perspective, efficient
distributed source coding always involves an error probability,even though it vanishes
as the coding block length is increased. In Chapter 4,Tuncel, Nayak,Koulgi, and Rose
take a more restrictive view: The error probability must be exactly zero.This is shown
to lead to a strict rate penalty for many instances. Chapter 5, by Goyal, Fletcher, and
Rangan, connects ideas from distributed source coding with the sparse signal models
that have recently received considerable attention under the heading of compressed
(or compressive) sensing.
The second part of the book focuses on algorithms and applications, where the
developments of the past decades have been even more pronounced than in the theoretical
foundations.The first chapter,by Guillemot and Roumy, presents an overview
of practical DSC techniques based on turbo and LDPC codes,along with ample experimental
illustration. Chapter 7, by Roy, Ajdler, Konsbruck, and Vetterli, specializes
and applies DSC techniques to a system of multiple microphones, using an explicit
spatial model to derive sampling conditions and source correlation structures. Chapter
8, by Pereira, Brites, and Ascenso, overviews the application of ideas from DSC
to video coding: A single video stream is encoded, frame by frame, and the encoder
treats past and future frames as side information when encoding the current frame.The
chapter starts with an overview of the original distributed video coders from Berkeley
(PRISM) and Stanford,and provides a detailed description of an enhanced video coder
developed by the authors (and referred to as DISCOVER). The case of the multiple
multiview video stream is considered by Nayak, Song, Tuncel, and Roy-Chowdhury
in Chapter 9,where they show how DSC techniques can be applied to the problem of
multiview video compression. Chapter 10, by Cheung and Ortega, applies DSC techniques
to the problem of distributed compression of hyperspectral imagery. Finally,
Chapter 11, by Vetro, Draper, Rane, and Yedidia, is an innovative application of DSC
techniques to securing biometric data.The problem is that if a fingerprint, iris scan,or
genetic code is used as a user password, then the password cannot be changed since
users are stuck with their fingers (or irises,or genes).Therefore,biometric information
should not be stored in the clear anywhere. This chapter discusses one approach to
this problematic issue, using ideas from DSC.
One of the main objectives of this book is to provide a comprehensive reference for
engineers, researchers, and students interested in distributed source coding. Results
on this topic have so far appeared in different journals and conferences. We hope
that the book will finally provide an integrated view of this active and ever evolving
research area.
Edited books would not exist without the enthusiasm and hard work of the
contributors. It has been a great pleasure for us to interact with some of the very
best researchers in this area who have enthusiastically embarked in this project
and have contributed these wonderful chapters. We have learned a lot from them.
We would also like to thank the reviewers of the chapters for their time and for
their constructive comments. Finally we would like to thank the staff at Academic
Press—in particularTim Pitts, Senior Commissioning Editor, and Melanie Benson—for
their continuous help.
Pier Luigi Dragotti, London, UK
Michael Gastpar, Berkeley, California, USA
本资源共7个可选网络硬盘链接，4.63 MB，保质期2009-07-15。
----------------------------------------------------------------------------
Distributed Source Coding_Theory, Algorithms and Applications.Pier Luigi Dragotti.Apress.2009.pdf
Distributed Source Coding_Theory, Algorithms and Applications.Pier Luigi Dragotti.Apress.2009.pdf
Distributed Source Coding_Theory, Algorithms and Applications.Pier Luigi Dragotti.Apress.2009.pdf
Distributed Source Coding_Theory, Algorithms and Applications.Pier Luigi Dragotti.Apress.2009.pdf
Distributed Source Coding_Theory, Algorithms and Applications.Pier Luigi Dragotti.Apress.2009.pdf
Distributed Source Coding_Theory, Algorithms and Applications.Pier Luigi Dragotti.Apress.2009.pdf
Distributed Source Coding_Theory, Algorithms and Applications.Pier Luigi Dragotti.Apress.2009.pdf
---------------------------------------------------------------------------- |
浙公网安备 33010202000686号 ( 浙ICP备09035230号-1 )
发表于 2009-9-23 09:42:37