Lossy Source Coding with Gaussian or Erased Side-Information

Suhas Diggavi, Etienne Perron, Emre Telatar
2009
Rapport ou document de travail

In this paper we find properties that are shared between two seemingly unrelated lossy source coding setups with side information. The first setup is when the source and side information are jointly Gaussian and the distortion measure is quadratic. The second setup is when the side information is an erased version of the source. We begin with the observation that in both these cases the Wyner-Ziv and conditional rate-distortion functions are equal. We further find that there is a continuum of optimal strategies for the conditional rate distortion problem in both these setups. Next, we consider the case when there are two decoders with access to different side-information sources. For the case when the encoder has access to the side information we establish bounds on the rate-distortion function and a sufficient condition for tightness. Under this condition, we find a characterization of the rate-distortion function for physically degraded side information. This characterization holds for both the Gaussian and erasure setups.

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Concepts associés

Chargement

Publications associées

Chargement

Suhas Diggavi, Etienne Perron, Emre Telatar
2009
Rapport ou document de travail

Résumé

Official source

https://infoscience.epfl.ch/record/131057?ln=fr

À propos de ce résultat

Concepts associés (7)

Concepts associés

Chargement

Publications associées (4)

Publications associées

Chargement

Lossy Source Coding with Gaussian or Erased Side-Information

Suhas Diggavi, Etienne Perron, Emre Telatar

In this paper we find properties that are shared between two seemingly unrelated lossy source coding setups with side-information. The first setup is when the source and side-information are jointly Gaussian and the distortion measure is quadratic. The second setup is when the side-information is an erased version of the source. We begin with the observation that in both these cases the Wyner-Ziv and conditional rate-distortion functions are equal. We further find that there is a continuum of optimal strategies for the conditional rate distortion problem in both these setups. Next, we consider the case when there are two decoders with access to different side-information sources. For the case when the encoder has access to the side-information we establish bounds on the rate-distortion function and a sufficient condition for tightness. Under this condition, we find a characterization of the rate-distortion function for physically degraded side-information. This characterization holds for both the Gaussian and erasure setups.

Ieee Service Center, 445 Hoes Lane, Po Box 1331, Piscataway, Nj 08855-1331 Usa2009

Chargement

The Kaspi Rate-Distortion Problem with Encoder Side-Information: Gaussian Case

Suhas Diggavi, Etienne Perron, Emre Telatar

We consider the problem of compressing a Gaussian source for two decoders, one of which has access to side-information. Kaspi has solved this problem for discrete memoryless sources for the two cases with and without encoder side-information. We focus on the case in which the encoder has access to side-information and we explicitly solve the optimization problem for Gaussian sources and squared error distortion measures. The achievability part of the proof provides an intuitive insight based on the idea of using innovations for the conditional rate-distortion problem. We compare our result to the case in which the encoder is not informed. Heegard and Berger have solved this problem for Gaussian sources. The comparison shows that in this multiple-decoder setup, side-information at the encoder improves the rate-distortion curve even in the Gaussian case.

2005

Chargement

On the Role of Encoder Side-Information in Source Coding for Multiple Decoders

Suhas Diggavi, Etienne Perron, Emre Telatar

We consider a lossy source coding problem where the description of a source is going to be used by two decoders, each having access to information correlated with the source. This side-information is also present at the encoder. We give inner and outer bounds to the set of achievable rate and distortion triples. For the special case of Gaussian sources with degraded side-information and squared error distortions, the two bounds coincide and we obtain the true rate-distortion region. As a further specialization, we obtain the rate-distortion region of the Gaussian version of a problem previously solved by Kaspi for discrete memoryless sources. Using this result, we quantify how much revealing the side-information to the encoder helps in such a Gaussian setup.

2006

Chargement

Rate–distortion theory

Rate–distortion theory is a major branch of information theory which provides the theoretical foundations for lossy data compression; it addresses the problem of determining the minimal number of bits

Asymétrie d'information

Une asymétrie d'information est une situation où, sur un marché, les agents économiques qui contractent ou échangent ne sont pas sur un pied d'égalité en termes d'informations, l'un des deux agents

Fonction gaussienne

vignette|Fonction gaussienne pour μ = 0, σ = 1 ; courbe centrée en zéro. Une fonction gaussienne est une fonction en exponentielle de l'opposé du carré de l'abscisse (une fonction en exp(-x)). Elle a