Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM

Yong Fan, Dinggang Shen, Christos Davatzikos

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

44 Citations (Scopus)

Abstract

This paper presents a method for classification of medical images, using machine learning and deformation-based morphometry. A morphological representation of the anatomy of interest is first obtained using high-dimensional template warping, from which regions that display strong correlations between morphological measurements and the classification (clinical) variable are extracted using a watershed segmentation, taking into account the regional smoothness of the correlation map which is estimated by a cross-validation strategy in order to achieve robustness to outliers. A Support Vector Machine-Recursive Feature Elimination (SVM-RFE) technique is then used to rank computed features from the extracted regions, according to their effect on the leave-one-out error bound. Finally, SVM classification is applied using the best set of features, and it is tested using leave-one-out. The results from a group of 61 brain images of female normal controls and schizophrenia patients demonstrate not only high classification accuracy (91.8%) and steep ROC curves, but also exceptional stability with respect to the number of selected features and the SVM kernel size.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages1-8
Number of pages8
Volume3749 LNCS
Publication statusPublished - 2005 Dec 1
Externally publishedYes
Event8th International Conference on Medical Image Computing and Computer-Assisted Intervention - MICCAI 2005 - Palm Springs, CA, United States
Duration: 2005 Oct 262005 Oct 29

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3749 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other8th International Conference on Medical Image Computing and Computer-Assisted Intervention - MICCAI 2005
CountryUnited States
CityPalm Springs, CA
Period05/10/2605/10/29

Fingerprint

Image Warping
Feature Extraction
Feature extraction
High-dimensional
Morphometry
Warping
Receiver Operating Characteristic Curve
Anatomy
Medical Image
Watersheds
Cross-validation
ROC Curve
Error Bounds
Outlier
Support vector machines
Learning systems
Elimination
Template
Smoothness
Brain

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Fan, Y., Shen, D., & Davatzikos, C. (2005). Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3749 LNCS, pp. 1-8). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3749 LNCS).

Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM. / Fan, Yong; Shen, Dinggang; Davatzikos, Christos.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3749 LNCS 2005. p. 1-8 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 3749 LNCS).

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

Fan, Y, Shen, D & Davatzikos, C 2005, Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 3749 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 3749 LNCS, pp. 1-8, 8th International Conference on Medical Image Computing and Computer-Assisted Intervention - MICCAI 2005, Palm Springs, CA, United States, 05/10/26.
Fan Y, Shen D, Davatzikos C. Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3749 LNCS. 2005. p. 1-8. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Fan, Yong ; Shen, Dinggang ; Davatzikos, Christos. / Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 3749 LNCS 2005. pp. 1-8 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{5acfa62c01b746118942eb14b647847d,
title = "Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM",
abstract = "This paper presents a method for classification of medical images, using machine learning and deformation-based morphometry. A morphological representation of the anatomy of interest is first obtained using high-dimensional template warping, from which regions that display strong correlations between morphological measurements and the classification (clinical) variable are extracted using a watershed segmentation, taking into account the regional smoothness of the correlation map which is estimated by a cross-validation strategy in order to achieve robustness to outliers. A Support Vector Machine-Recursive Feature Elimination (SVM-RFE) technique is then used to rank computed features from the extracted regions, according to their effect on the leave-one-out error bound. Finally, SVM classification is applied using the best set of features, and it is tested using leave-one-out. The results from a group of 61 brain images of female normal controls and schizophrenia patients demonstrate not only high classification accuracy (91.8{\%}) and steep ROC curves, but also exceptional stability with respect to the number of selected features and the SVM kernel size.",
author = "Yong Fan and Dinggang Shen and Christos Davatzikos",
year = "2005",
month = "12",
day = "1",
language = "English",
isbn = "3540293272",
volume = "3749 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "1--8",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM

AU - Fan, Yong

AU - Shen, Dinggang

AU - Davatzikos, Christos

PY - 2005/12/1

Y1 - 2005/12/1

N2 - This paper presents a method for classification of medical images, using machine learning and deformation-based morphometry. A morphological representation of the anatomy of interest is first obtained using high-dimensional template warping, from which regions that display strong correlations between morphological measurements and the classification (clinical) variable are extracted using a watershed segmentation, taking into account the regional smoothness of the correlation map which is estimated by a cross-validation strategy in order to achieve robustness to outliers. A Support Vector Machine-Recursive Feature Elimination (SVM-RFE) technique is then used to rank computed features from the extracted regions, according to their effect on the leave-one-out error bound. Finally, SVM classification is applied using the best set of features, and it is tested using leave-one-out. The results from a group of 61 brain images of female normal controls and schizophrenia patients demonstrate not only high classification accuracy (91.8%) and steep ROC curves, but also exceptional stability with respect to the number of selected features and the SVM kernel size.

AB - This paper presents a method for classification of medical images, using machine learning and deformation-based morphometry. A morphological representation of the anatomy of interest is first obtained using high-dimensional template warping, from which regions that display strong correlations between morphological measurements and the classification (clinical) variable are extracted using a watershed segmentation, taking into account the regional smoothness of the correlation map which is estimated by a cross-validation strategy in order to achieve robustness to outliers. A Support Vector Machine-Recursive Feature Elimination (SVM-RFE) technique is then used to rank computed features from the extracted regions, according to their effect on the leave-one-out error bound. Finally, SVM classification is applied using the best set of features, and it is tested using leave-one-out. The results from a group of 61 brain images of female normal controls and schizophrenia patients demonstrate not only high classification accuracy (91.8%) and steep ROC curves, but also exceptional stability with respect to the number of selected features and the SVM kernel size.

UR - http://www.scopus.com/inward/record.url?scp=33646852713&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33646852713&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:33646852713

SN - 3540293272

SN - 9783540293279

VL - 3749 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1

EP - 8

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -