Wavelet-based compressed sensing using a gaussian scale mixture model

Yookyung Kim, Mariappan S. Nadar, Ali Bilgin

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

While initial compressed sensing (CS) recovery techniques operated under the implicit assumption that the sparse domain coefficients are independently distributed, recent results have indicated that integrating a statistical or structural dependence model of sparse domain coefficients into CS enhances recovery. In this paper, we present a method for exploiting empirical dependences among wavelet coefficients during CS recovery using a Bayes least-square Gaussian-scale-mixture model. The proposed model is successfully incorporated into several recent CS algorithms, including reweighted l 1 minimization (RL1), iteratively reweighted least squares, and iterative hard thresholding. Extensive experiments including comparisons with a state-of-the-art model-based CS method demonstrate that the proposed algorithms are highly effective at reducing reconstruction error and/or the number of measurements required for a desired reconstruction quality.

Original languageEnglish (US)
Article number6156783
Pages (from-to)3102-3108
Number of pages7
JournalIEEE Transactions on Image Processing
Volume21
Issue number6
DOIs
StatePublished - Jun 2012

Fingerprint

Compressed sensing
Recovery
Experiments

Keywords

  • Compressed sensing (CS)
  • Gaussian scale mixtures (GSMs)
  • Structured sparsity

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Software

Cite this

Wavelet-based compressed sensing using a gaussian scale mixture model. / Kim, Yookyung; Nadar, Mariappan S.; Bilgin, Ali.

In: IEEE Transactions on Image Processing, Vol. 21, No. 6, 6156783, 06.2012, p. 3102-3108.

Research output: Contribution to journalArticle

Kim, Yookyung ; Nadar, Mariappan S. ; Bilgin, Ali. / Wavelet-based compressed sensing using a gaussian scale mixture model. In: IEEE Transactions on Image Processing. 2012 ; Vol. 21, No. 6. pp. 3102-3108.
@article{9660327ac73a4fbc903b831e28e03047,
title = "Wavelet-based compressed sensing using a gaussian scale mixture model",
abstract = "While initial compressed sensing (CS) recovery techniques operated under the implicit assumption that the sparse domain coefficients are independently distributed, recent results have indicated that integrating a statistical or structural dependence model of sparse domain coefficients into CS enhances recovery. In this paper, we present a method for exploiting empirical dependences among wavelet coefficients during CS recovery using a Bayes least-square Gaussian-scale-mixture model. The proposed model is successfully incorporated into several recent CS algorithms, including reweighted l 1 minimization (RL1), iteratively reweighted least squares, and iterative hard thresholding. Extensive experiments including comparisons with a state-of-the-art model-based CS method demonstrate that the proposed algorithms are highly effective at reducing reconstruction error and/or the number of measurements required for a desired reconstruction quality.",
keywords = "Compressed sensing (CS), Gaussian scale mixtures (GSMs), Structured sparsity",
author = "Yookyung Kim and Nadar, {Mariappan S.} and Ali Bilgin",
year = "2012",
month = "6",
doi = "10.1109/TIP.2012.2188807",
language = "English (US)",
volume = "21",
pages = "3102--3108",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

TY - JOUR

T1 - Wavelet-based compressed sensing using a gaussian scale mixture model

AU - Kim, Yookyung

AU - Nadar, Mariappan S.

AU - Bilgin, Ali

PY - 2012/6

Y1 - 2012/6

N2 - While initial compressed sensing (CS) recovery techniques operated under the implicit assumption that the sparse domain coefficients are independently distributed, recent results have indicated that integrating a statistical or structural dependence model of sparse domain coefficients into CS enhances recovery. In this paper, we present a method for exploiting empirical dependences among wavelet coefficients during CS recovery using a Bayes least-square Gaussian-scale-mixture model. The proposed model is successfully incorporated into several recent CS algorithms, including reweighted l 1 minimization (RL1), iteratively reweighted least squares, and iterative hard thresholding. Extensive experiments including comparisons with a state-of-the-art model-based CS method demonstrate that the proposed algorithms are highly effective at reducing reconstruction error and/or the number of measurements required for a desired reconstruction quality.

AB - While initial compressed sensing (CS) recovery techniques operated under the implicit assumption that the sparse domain coefficients are independently distributed, recent results have indicated that integrating a statistical or structural dependence model of sparse domain coefficients into CS enhances recovery. In this paper, we present a method for exploiting empirical dependences among wavelet coefficients during CS recovery using a Bayes least-square Gaussian-scale-mixture model. The proposed model is successfully incorporated into several recent CS algorithms, including reweighted l 1 minimization (RL1), iteratively reweighted least squares, and iterative hard thresholding. Extensive experiments including comparisons with a state-of-the-art model-based CS method demonstrate that the proposed algorithms are highly effective at reducing reconstruction error and/or the number of measurements required for a desired reconstruction quality.

KW - Compressed sensing (CS)

KW - Gaussian scale mixtures (GSMs)

KW - Structured sparsity

UR - http://www.scopus.com/inward/record.url?scp=84861108626&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84861108626&partnerID=8YFLogxK

U2 - 10.1109/TIP.2012.2188807

DO - 10.1109/TIP.2012.2188807

M3 - Article

VL - 21

SP - 3102

EP - 3108

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

IS - 6

M1 - 6156783

ER -