Estimation Risk of Transformation-Averaged Estimators

Juan Liu, Pierre Moulin

Research output: Contribution to journalConference article

Abstract

Wavelet image denoising practice has shown that the performance of simple estimators may be substantially improved by averaging these estimators over a collection of transformations such as translations or rotations. In this paper, we explain and quantify these empirical findings using estimation theory. We consider a general nonlinear observation model, analyze the estimation risk of transformation-averaged estimators, and derive an upper bound on the risk reduction due to transformation averaging. The bound is evaluated for several estimators, using different averaging strategies (including a randomized strategy) and different wavelet bases. The practical usefulness of the bound is established for standard image denoising examples.

Original languageEnglish (US)
Pages (from-to)87-96
Number of pages10
JournalProceedings of SPIE - The International Society for Optical Engineering
Volume5207
Issue number1
StatePublished - Dec 1 2003
EventWavelets: Applications in Signal and Image Processing X - San Diego, CA, United States
Duration: Aug 4 2003Aug 8 2003

Fingerprint

Image denoising
estimators
Averaging
Estimator
Image Denoising
Wavelet Denoising
Estimation Theory
Wavelet Bases
Quantify
Upper bound
Strategy

Keywords

  • Estimation risk
  • Invariance
  • Transformation groups
  • Wavelets

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this

Estimation Risk of Transformation-Averaged Estimators. / Liu, Juan; Moulin, Pierre.

In: Proceedings of SPIE - The International Society for Optical Engineering, Vol. 5207, No. 1, 01.12.2003, p. 87-96.

Research output: Contribution to journalConference article

@article{d8ee130753a145f2bb46f1624e4b6926,
title = "Estimation Risk of Transformation-Averaged Estimators",
abstract = "Wavelet image denoising practice has shown that the performance of simple estimators may be substantially improved by averaging these estimators over a collection of transformations such as translations or rotations. In this paper, we explain and quantify these empirical findings using estimation theory. We consider a general nonlinear observation model, analyze the estimation risk of transformation-averaged estimators, and derive an upper bound on the risk reduction due to transformation averaging. The bound is evaluated for several estimators, using different averaging strategies (including a randomized strategy) and different wavelet bases. The practical usefulness of the bound is established for standard image denoising examples.",
keywords = "Estimation risk, Invariance, Transformation groups, Wavelets",
author = "Juan Liu and Pierre Moulin",
year = "2003",
month = "12",
day = "1",
language = "English (US)",
volume = "5207",
pages = "87--96",
journal = "Proceedings of SPIE - The International Society for Optical Engineering",
issn = "0277-786X",
publisher = "SPIE",
number = "1",

}

TY - JOUR

T1 - Estimation Risk of Transformation-Averaged Estimators

AU - Liu, Juan

AU - Moulin, Pierre

PY - 2003/12/1

Y1 - 2003/12/1

N2 - Wavelet image denoising practice has shown that the performance of simple estimators may be substantially improved by averaging these estimators over a collection of transformations such as translations or rotations. In this paper, we explain and quantify these empirical findings using estimation theory. We consider a general nonlinear observation model, analyze the estimation risk of transformation-averaged estimators, and derive an upper bound on the risk reduction due to transformation averaging. The bound is evaluated for several estimators, using different averaging strategies (including a randomized strategy) and different wavelet bases. The practical usefulness of the bound is established for standard image denoising examples.

AB - Wavelet image denoising practice has shown that the performance of simple estimators may be substantially improved by averaging these estimators over a collection of transformations such as translations or rotations. In this paper, we explain and quantify these empirical findings using estimation theory. We consider a general nonlinear observation model, analyze the estimation risk of transformation-averaged estimators, and derive an upper bound on the risk reduction due to transformation averaging. The bound is evaluated for several estimators, using different averaging strategies (including a randomized strategy) and different wavelet bases. The practical usefulness of the bound is established for standard image denoising examples.

KW - Estimation risk

KW - Invariance

KW - Transformation groups

KW - Wavelets

UR - http://www.scopus.com/inward/record.url?scp=1242286066&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=1242286066&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:1242286066

VL - 5207

SP - 87

EP - 96

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

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

SN - 0277-786X

IS - 1

ER -