Wavelet and curvelet moments for image classification

Application to aggregate mixture grading

Fionn Murtagh, Jean Luc Starck

Research output: Contribution to journalArticle

47 Citations (Scopus)

Abstract

We show the potential for classifying images of mixtures of aggregate, based themselves on varying, albeit well-defined, sizes and shapes, in order to provide a far more effective approach compared to the classification of individual sizes and shapes. While a dominant (additive, stationary) Gaussian noise component in image data will ensure that wavelet coefficients are of Gaussian distribution, long tailed distributions (symptomatic, for example, of extreme values) may well hold in practice for wavelet coefficients. Energy (second order moment) has often been used for image characterization for image content-based retrieval, and higher order moments may be important also, not least for capturing long tailed distributional behavior. In this work, we assess second, third and fourth order moments of multiresolution transform - wavelet and curvelet transform - coefficients as features. As analysis methodology, taking account of image types, multiresolution transforms, and moments of coefficients in the scales or bands, we use correspondence analysis as well as k-nearest neighbors supervised classification.

Original languageEnglish
Pages (from-to)1557-1564
Number of pages8
JournalPattern Recognition Letters
Volume29
Issue number10
DOIs
Publication statusPublished - 15 Jul 2008
Externally publishedYes

Fingerprint

Image classification
Content based retrieval
Gaussian distribution
Wavelet transforms

Cite this

@article{c276b515c19d43e0b73c91602eed7c33,
title = "Wavelet and curvelet moments for image classification: Application to aggregate mixture grading",
abstract = "We show the potential for classifying images of mixtures of aggregate, based themselves on varying, albeit well-defined, sizes and shapes, in order to provide a far more effective approach compared to the classification of individual sizes and shapes. While a dominant (additive, stationary) Gaussian noise component in image data will ensure that wavelet coefficients are of Gaussian distribution, long tailed distributions (symptomatic, for example, of extreme values) may well hold in practice for wavelet coefficients. Energy (second order moment) has often been used for image characterization for image content-based retrieval, and higher order moments may be important also, not least for capturing long tailed distributional behavior. In this work, we assess second, third and fourth order moments of multiresolution transform - wavelet and curvelet transform - coefficients as features. As analysis methodology, taking account of image types, multiresolution transforms, and moments of coefficients in the scales or bands, we use correspondence analysis as well as k-nearest neighbors supervised classification.",
keywords = "Image grading, Kurtosis, Moments, Skewness, Variance, Wavelet and curvelet transforms",
author = "Fionn Murtagh and Starck, {Jean Luc}",
year = "2008",
month = "7",
day = "15",
doi = "10.1016/j.patrec.2008.03.008",
language = "English",
volume = "29",
pages = "1557--1564",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "North-Holland Publ Co",
number = "10",

}

Wavelet and curvelet moments for image classification : Application to aggregate mixture grading. / Murtagh, Fionn; Starck, Jean Luc.

In: Pattern Recognition Letters, Vol. 29, No. 10, 15.07.2008, p. 1557-1564.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Wavelet and curvelet moments for image classification

T2 - Application to aggregate mixture grading

AU - Murtagh, Fionn

AU - Starck, Jean Luc

PY - 2008/7/15

Y1 - 2008/7/15

N2 - We show the potential for classifying images of mixtures of aggregate, based themselves on varying, albeit well-defined, sizes and shapes, in order to provide a far more effective approach compared to the classification of individual sizes and shapes. While a dominant (additive, stationary) Gaussian noise component in image data will ensure that wavelet coefficients are of Gaussian distribution, long tailed distributions (symptomatic, for example, of extreme values) may well hold in practice for wavelet coefficients. Energy (second order moment) has often been used for image characterization for image content-based retrieval, and higher order moments may be important also, not least for capturing long tailed distributional behavior. In this work, we assess second, third and fourth order moments of multiresolution transform - wavelet and curvelet transform - coefficients as features. As analysis methodology, taking account of image types, multiresolution transforms, and moments of coefficients in the scales or bands, we use correspondence analysis as well as k-nearest neighbors supervised classification.

AB - We show the potential for classifying images of mixtures of aggregate, based themselves on varying, albeit well-defined, sizes and shapes, in order to provide a far more effective approach compared to the classification of individual sizes and shapes. While a dominant (additive, stationary) Gaussian noise component in image data will ensure that wavelet coefficients are of Gaussian distribution, long tailed distributions (symptomatic, for example, of extreme values) may well hold in practice for wavelet coefficients. Energy (second order moment) has often been used for image characterization for image content-based retrieval, and higher order moments may be important also, not least for capturing long tailed distributional behavior. In this work, we assess second, third and fourth order moments of multiresolution transform - wavelet and curvelet transform - coefficients as features. As analysis methodology, taking account of image types, multiresolution transforms, and moments of coefficients in the scales or bands, we use correspondence analysis as well as k-nearest neighbors supervised classification.

KW - Image grading

KW - Kurtosis

KW - Moments

KW - Skewness

KW - Variance

KW - Wavelet and curvelet transforms

UR - http://www.scopus.com/inward/record.url?scp=44649193569&partnerID=8YFLogxK

U2 - 10.1016/j.patrec.2008.03.008

DO - 10.1016/j.patrec.2008.03.008

M3 - Article

VL - 29

SP - 1557

EP - 1564

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

IS - 10

ER -