Image denoising using a local Gaussian scale mixture model in the wavelet domain

Vasily Strela, Javier Portilla, Eero Simoncelli

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The statistics of photographic images, when decomposed in a multiscale wavelet basis, exhibit striking non-Gaussian behaviors. The joint densities of clusters of wavelet coefficients (corresponding to basis functions at nearby spatial positions, orientations and scales) are well-described as a Gaussian scale mixture: a jointly Gaussian vector multiplied by a hidden scaling variable. We develop a maximum likelihood solution for estimating the hidden variable from an observation of the cluster of coefficients contaminated by additive Gaussian noise. The estimated hidden variable is then used to estimate the original noise-free coefficients. We demonstrate the power of this model through numerical simulations of image denoising.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
Pages363-371
Number of pages9
Volume4119
DOIs
StatePublished - 2000
EventWavelet Applications in Signal and Image Processing VIII - San Diego, CA, USA
Duration: Jul 31 2000Aug 4 2000

Other

OtherWavelet Applications in Signal and Image Processing VIII
CitySan Diego, CA, USA
Period7/31/008/4/00

Fingerprint

Image denoising
Maximum likelihood
Statistics
Computer simulation
coefficients
random noise
estimating
statistics
scaling
estimates
simulation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Strela, V., Portilla, J., & Simoncelli, E. (2000). Image denoising using a local Gaussian scale mixture model in the wavelet domain. In Proceedings of SPIE - The International Society for Optical Engineering (Vol. 4119, pp. 363-371) https://doi.org/10.1117/12.408621

Image denoising using a local Gaussian scale mixture model in the wavelet domain. / Strela, Vasily; Portilla, Javier; Simoncelli, Eero.

Proceedings of SPIE - The International Society for Optical Engineering. Vol. 4119 2000. p. 363-371.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Strela, V, Portilla, J & Simoncelli, E 2000, Image denoising using a local Gaussian scale mixture model in the wavelet domain. in Proceedings of SPIE - The International Society for Optical Engineering. vol. 4119, pp. 363-371, Wavelet Applications in Signal and Image Processing VIII, San Diego, CA, USA, 7/31/00. https://doi.org/10.1117/12.408621
Strela V, Portilla J, Simoncelli E. Image denoising using a local Gaussian scale mixture model in the wavelet domain. In Proceedings of SPIE - The International Society for Optical Engineering. Vol. 4119. 2000. p. 363-371 https://doi.org/10.1117/12.408621
Strela, Vasily ; Portilla, Javier ; Simoncelli, Eero. / Image denoising using a local Gaussian scale mixture model in the wavelet domain. Proceedings of SPIE - The International Society for Optical Engineering. Vol. 4119 2000. pp. 363-371
@inproceedings{346d3ba622bd4d2f8940686c91b64c2a,
title = "Image denoising using a local Gaussian scale mixture model in the wavelet domain",
abstract = "The statistics of photographic images, when decomposed in a multiscale wavelet basis, exhibit striking non-Gaussian behaviors. The joint densities of clusters of wavelet coefficients (corresponding to basis functions at nearby spatial positions, orientations and scales) are well-described as a Gaussian scale mixture: a jointly Gaussian vector multiplied by a hidden scaling variable. We develop a maximum likelihood solution for estimating the hidden variable from an observation of the cluster of coefficients contaminated by additive Gaussian noise. The estimated hidden variable is then used to estimate the original noise-free coefficients. We demonstrate the power of this model through numerical simulations of image denoising.",
author = "Vasily Strela and Javier Portilla and Eero Simoncelli",
year = "2000",
doi = "10.1117/12.408621",
language = "English (US)",
volume = "4119",
pages = "363--371",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

}

TY - GEN

T1 - Image denoising using a local Gaussian scale mixture model in the wavelet domain

AU - Strela, Vasily

AU - Portilla, Javier

AU - Simoncelli, Eero

PY - 2000

Y1 - 2000

N2 - The statistics of photographic images, when decomposed in a multiscale wavelet basis, exhibit striking non-Gaussian behaviors. The joint densities of clusters of wavelet coefficients (corresponding to basis functions at nearby spatial positions, orientations and scales) are well-described as a Gaussian scale mixture: a jointly Gaussian vector multiplied by a hidden scaling variable. We develop a maximum likelihood solution for estimating the hidden variable from an observation of the cluster of coefficients contaminated by additive Gaussian noise. The estimated hidden variable is then used to estimate the original noise-free coefficients. We demonstrate the power of this model through numerical simulations of image denoising.

AB - The statistics of photographic images, when decomposed in a multiscale wavelet basis, exhibit striking non-Gaussian behaviors. The joint densities of clusters of wavelet coefficients (corresponding to basis functions at nearby spatial positions, orientations and scales) are well-described as a Gaussian scale mixture: a jointly Gaussian vector multiplied by a hidden scaling variable. We develop a maximum likelihood solution for estimating the hidden variable from an observation of the cluster of coefficients contaminated by additive Gaussian noise. The estimated hidden variable is then used to estimate the original noise-free coefficients. We demonstrate the power of this model through numerical simulations of image denoising.

UR - http://www.scopus.com/inward/record.url?scp=0034430087&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034430087&partnerID=8YFLogxK

U2 - 10.1117/12.408621

DO - 10.1117/12.408621

M3 - Conference contribution

VL - 4119

SP - 363

EP - 371

BT - Proceedings of SPIE - The International Society for Optical Engineering

ER -