Comparison of traditional brain segmentation tools with 3D self-organizing maps

David Dean, Krishnamurthy Subramanyan, Janardhan Kamath, Fred Bookstein, David Wilson, David Kwon, Peter Buckley

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

3 Citations (Scopus)

Abstract

Algorithm-assisted 3D MR brain segmentation may be significantly faster than manual methods and produce visually pleasing results. We tested two- and three-dimensional region growing (2DRG and 3DRG) and selforganizing map (SOM) algorithms for segmentation of the cerebral ventricles. The SOM algorithm provides the greatest times savings, 12:1, over manual segmentation. Concern for reproducibility of algorithm-assisted segmentation motivated an intra-operator comparative study of these and manual segmentation methods. One of us, DK, segmented the cerebral ventricles from 5 3D MR-scan data sets three times manually and with the three algorithms. When variability is measured as the shape variance of derived landmarks sets, the three algorithm-assisted methods show less intra-operator variability than manual segmentation. The 2DRG and 3DRG segmentations show more variability than SOM. Of the 4 methods, SOM segmentation requires the fewest operator decisions.

Original languageEnglish (US)
Title of host publicationInformation Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings
EditorsGene Gindi, James Duncan
PublisherSpringer Verlag
Pages393-398
Number of pages6
ISBN (Print)3540630465, 9783540630463
StatePublished - Jan 1 1997
Event15th International Conference on Information Processing in Medical Imaging, IPMI 1997 - Poultney, United States
Duration: Jun 9 1997Jun 13 1997

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1230
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other15th International Conference on Information Processing in Medical Imaging, IPMI 1997
CountryUnited States
CityPoultney
Period6/9/976/13/97

Fingerprint

Self organizing maps
Self-organizing Map
Brain
Segmentation
Mathematical operators
Operator
Region Growing
Reproducibility
Landmarks
Comparative Study
Three-dimensional

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Dean, D., Subramanyan, K., Kamath, J., Bookstein, F., Wilson, D., Kwon, D., & Buckley, P. (1997). Comparison of traditional brain segmentation tools with 3D self-organizing maps. In G. Gindi, & J. Duncan (Eds.), Information Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings (pp. 393-398). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1230). Springer Verlag.

Comparison of traditional brain segmentation tools with 3D self-organizing maps. / Dean, David; Subramanyan, Krishnamurthy; Kamath, Janardhan; Bookstein, Fred; Wilson, David; Kwon, David; Buckley, Peter.

Information Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings. ed. / Gene Gindi; James Duncan. Springer Verlag, 1997. p. 393-398 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1230).

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

Dean, D, Subramanyan, K, Kamath, J, Bookstein, F, Wilson, D, Kwon, D & Buckley, P 1997, Comparison of traditional brain segmentation tools with 3D self-organizing maps. in G Gindi & J Duncan (eds), Information Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1230, Springer Verlag, pp. 393-398, 15th International Conference on Information Processing in Medical Imaging, IPMI 1997, Poultney, United States, 6/9/97.
Dean D, Subramanyan K, Kamath J, Bookstein F, Wilson D, Kwon D et al. Comparison of traditional brain segmentation tools with 3D self-organizing maps. In Gindi G, Duncan J, editors, Information Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings. Springer Verlag. 1997. p. 393-398. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Dean, David ; Subramanyan, Krishnamurthy ; Kamath, Janardhan ; Bookstein, Fred ; Wilson, David ; Kwon, David ; Buckley, Peter. / Comparison of traditional brain segmentation tools with 3D self-organizing maps. Information Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings. editor / Gene Gindi ; James Duncan. Springer Verlag, 1997. pp. 393-398 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{f7411d20fb284775a6fc8f98f747f64a,
title = "Comparison of traditional brain segmentation tools with 3D self-organizing maps",
abstract = "Algorithm-assisted 3D MR brain segmentation may be significantly faster than manual methods and produce visually pleasing results. We tested two- and three-dimensional region growing (2DRG and 3DRG) and selforganizing map (SOM) algorithms for segmentation of the cerebral ventricles. The SOM algorithm provides the greatest times savings, 12:1, over manual segmentation. Concern for reproducibility of algorithm-assisted segmentation motivated an intra-operator comparative study of these and manual segmentation methods. One of us, DK, segmented the cerebral ventricles from 5 3D MR-scan data sets three times manually and with the three algorithms. When variability is measured as the shape variance of derived landmarks sets, the three algorithm-assisted methods show less intra-operator variability than manual segmentation. The 2DRG and 3DRG segmentations show more variability than SOM. Of the 4 methods, SOM segmentation requires the fewest operator decisions.",
author = "David Dean and Krishnamurthy Subramanyan and Janardhan Kamath and Fred Bookstein and David Wilson and David Kwon and Peter Buckley",
year = "1997",
month = "1",
day = "1",
language = "English (US)",
isbn = "3540630465",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "393--398",
editor = "Gene Gindi and James Duncan",
booktitle = "Information Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings",

}

TY - GEN

T1 - Comparison of traditional brain segmentation tools with 3D self-organizing maps

AU - Dean, David

AU - Subramanyan, Krishnamurthy

AU - Kamath, Janardhan

AU - Bookstein, Fred

AU - Wilson, David

AU - Kwon, David

AU - Buckley, Peter

PY - 1997/1/1

Y1 - 1997/1/1

N2 - Algorithm-assisted 3D MR brain segmentation may be significantly faster than manual methods and produce visually pleasing results. We tested two- and three-dimensional region growing (2DRG and 3DRG) and selforganizing map (SOM) algorithms for segmentation of the cerebral ventricles. The SOM algorithm provides the greatest times savings, 12:1, over manual segmentation. Concern for reproducibility of algorithm-assisted segmentation motivated an intra-operator comparative study of these and manual segmentation methods. One of us, DK, segmented the cerebral ventricles from 5 3D MR-scan data sets three times manually and with the three algorithms. When variability is measured as the shape variance of derived landmarks sets, the three algorithm-assisted methods show less intra-operator variability than manual segmentation. The 2DRG and 3DRG segmentations show more variability than SOM. Of the 4 methods, SOM segmentation requires the fewest operator decisions.

AB - Algorithm-assisted 3D MR brain segmentation may be significantly faster than manual methods and produce visually pleasing results. We tested two- and three-dimensional region growing (2DRG and 3DRG) and selforganizing map (SOM) algorithms for segmentation of the cerebral ventricles. The SOM algorithm provides the greatest times savings, 12:1, over manual segmentation. Concern for reproducibility of algorithm-assisted segmentation motivated an intra-operator comparative study of these and manual segmentation methods. One of us, DK, segmented the cerebral ventricles from 5 3D MR-scan data sets three times manually and with the three algorithms. When variability is measured as the shape variance of derived landmarks sets, the three algorithm-assisted methods show less intra-operator variability than manual segmentation. The 2DRG and 3DRG segmentations show more variability than SOM. Of the 4 methods, SOM segmentation requires the fewest operator decisions.

UR - http://www.scopus.com/inward/record.url?scp=21744439205&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=21744439205&partnerID=8YFLogxK

M3 - Conference contribution

SN - 3540630465

SN - 9783540630463

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

SP - 393

EP - 398

BT - Information Processing in Medical Imaging - 15th International Conference, IPMI 1997, Proceedings

A2 - Gindi, Gene

A2 - Duncan, James

PB - Springer Verlag

ER -