Vector Gaussian multiple description with individual and central receivers

Research output: Contribution to journalArticle

Abstract

L multiple descriptions of a vector Gaussian source for individual and central receivers are investigated. The sum rate of the descriptions with covariance distortion measure constraints, in a positive semidefinite ordering, is exactly characterized. For two descriptions, the entire rate region is characterized. The key component of the solution is a novel information-theoretic inequality that is used to lower-bound the achievable multiple description rates. Jointly Gaussian descriptions are optimal in achieving the limiting rates. We also show the robustness of this description scheme: The distortions achieved are no larger when used to describe any non-Gaussian source with the same covariance matrix.

Original languageEnglish (US)
Pages (from-to)2133-2153
Number of pages21
JournalIEEE Transactions on Information Theory
Volume53
Issue number6
DOIs
StatePublished - Jun 1 2007

Fingerprint

recipient
Covariance matrix

Keywords

  • Lossy compression
  • Multiple description
  • Quadratic distortion

ASJC Scopus subject areas

  • Information Systems
  • Computer Science Applications
  • Library and Information Sciences

Cite this

Vector Gaussian multiple description with individual and central receivers. / Wang, Hua; Viswanath, Pramod.

In: IEEE Transactions on Information Theory, Vol. 53, No. 6, 01.06.2007, p. 2133-2153.

Research output: Contribution to journalArticle

@article{9068897c08eb497197d3e30d1b0d2ed2,
title = "Vector Gaussian multiple description with individual and central receivers",
abstract = "L multiple descriptions of a vector Gaussian source for individual and central receivers are investigated. The sum rate of the descriptions with covariance distortion measure constraints, in a positive semidefinite ordering, is exactly characterized. For two descriptions, the entire rate region is characterized. The key component of the solution is a novel information-theoretic inequality that is used to lower-bound the achievable multiple description rates. Jointly Gaussian descriptions are optimal in achieving the limiting rates. We also show the robustness of this description scheme: The distortions achieved are no larger when used to describe any non-Gaussian source with the same covariance matrix.",
keywords = "Lossy compression, Multiple description, Quadratic distortion",
author = "Hua Wang and Pramod Viswanath",
year = "2007",
month = "6",
day = "1",
doi = "10.1109/TIT.2007.896880",
language = "English (US)",
volume = "53",
pages = "2133--2153",
journal = "IEEE Transactions on Information Theory",
issn = "0018-9448",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

TY - JOUR

T1 - Vector Gaussian multiple description with individual and central receivers

AU - Wang, Hua

AU - Viswanath, Pramod

PY - 2007/6/1

Y1 - 2007/6/1

N2 - L multiple descriptions of a vector Gaussian source for individual and central receivers are investigated. The sum rate of the descriptions with covariance distortion measure constraints, in a positive semidefinite ordering, is exactly characterized. For two descriptions, the entire rate region is characterized. The key component of the solution is a novel information-theoretic inequality that is used to lower-bound the achievable multiple description rates. Jointly Gaussian descriptions are optimal in achieving the limiting rates. We also show the robustness of this description scheme: The distortions achieved are no larger when used to describe any non-Gaussian source with the same covariance matrix.

AB - L multiple descriptions of a vector Gaussian source for individual and central receivers are investigated. The sum rate of the descriptions with covariance distortion measure constraints, in a positive semidefinite ordering, is exactly characterized. For two descriptions, the entire rate region is characterized. The key component of the solution is a novel information-theoretic inequality that is used to lower-bound the achievable multiple description rates. Jointly Gaussian descriptions are optimal in achieving the limiting rates. We also show the robustness of this description scheme: The distortions achieved are no larger when used to describe any non-Gaussian source with the same covariance matrix.

KW - Lossy compression

KW - Multiple description

KW - Quadratic distortion

UR - http://www.scopus.com/inward/record.url?scp=34249786865&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34249786865&partnerID=8YFLogxK

U2 - 10.1109/TIT.2007.896880

DO - 10.1109/TIT.2007.896880

M3 - Article

AN - SCOPUS:34249786865

VL - 53

SP - 2133

EP - 2153

JO - IEEE Transactions on Information Theory

JF - IEEE Transactions on Information Theory

SN - 0018-9448

IS - 6

ER -