Asymmetric similarity-weighted ensembles for image segmentation

V. Cheplygina, A. Van Opbroek, M.A. Ikram, Meike W. Vernooij, M. De Bruijne

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

Abstract

Supervised classification is widely used for image segmentation. To work effectively, these techniques need large amounts of labeled training data, that is representative of the test data. Different patient groups, different scanners or different scanning protocols can lead to differences between the images, thus representative data might not be available. Transfer learning techniques can be used to account for these differences, thus taking advantage of all the available data acquired with different protocols. We investigate the use of classifier ensembles, where each classifier is weighted according to the similarity between the data it is trained on, and the data it needs to segment. We examine 3 asymmetric similarity measures that can be used in scenarios where no labeled data from a newly introduced scanner or scanning protocol is available. We show that the asymmetry is informative and the direction of measurement needs to be chosen carefully. We also show that a point set similarity measure is robust across different studies, and outperforms state-of-the-art results on a multi-center brain tissue segmentation task.

Original languageEnglish
Title of host publication2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings
PublisherIEEE Computer Society
Pages273-277
Number of pages5
Volume2016-June
ISBN (Electronic)9781479923502
DOIs
Publication statusPublished - 15 Jun 2016
Event2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Prague, Czech Republic
Duration: 13 Apr 201616 Apr 2016

Conference

Conference2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016
CountryCzech Republic
CityPrague
Period13/04/201616/04/2016

Cite this

Cheplygina, V., Van Opbroek, A., Ikram, M. A., Vernooij, M. W., & De Bruijne, M. (2016). Asymmetric similarity-weighted ensembles for image segmentation. In 2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings (Vol. 2016-June, pp. 273-277). [7493262] IEEE Computer Society. https://doi.org/10.1109/ISBI.2016.7493262
Cheplygina, V. ; Van Opbroek, A. ; Ikram, M.A. ; Vernooij, Meike W. ; De Bruijne, M. / Asymmetric similarity-weighted ensembles for image segmentation. 2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings. Vol. 2016-June IEEE Computer Society, 2016. pp. 273-277
@inproceedings{26c82d5ce10b47d0a18e763ba3fa806c,
title = "Asymmetric similarity-weighted ensembles for image segmentation",
abstract = "Supervised classification is widely used for image segmentation. To work effectively, these techniques need large amounts of labeled training data, that is representative of the test data. Different patient groups, different scanners or different scanning protocols can lead to differences between the images, thus representative data might not be available. Transfer learning techniques can be used to account for these differences, thus taking advantage of all the available data acquired with different protocols. We investigate the use of classifier ensembles, where each classifier is weighted according to the similarity between the data it is trained on, and the data it needs to segment. We examine 3 asymmetric similarity measures that can be used in scenarios where no labeled data from a newly introduced scanner or scanning protocol is available. We show that the asymmetry is informative and the direction of measurement needs to be chosen carefully. We also show that a point set similarity measure is robust across different studies, and outperforms state-of-the-art results on a multi-center brain tissue segmentation task.",
keywords = "asymmetry, similarity measure, tissue segmentation, Transfer learning",
author = "V. Cheplygina and {Van Opbroek}, A. and M.A. Ikram and Vernooij, {Meike W.} and {De Bruijne}, M.",
year = "2016",
month = "6",
day = "15",
doi = "10.1109/ISBI.2016.7493262",
language = "English",
volume = "2016-June",
pages = "273--277",
booktitle = "2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings",
publisher = "IEEE Computer Society",
address = "United States",

}

Cheplygina, V, Van Opbroek, A, Ikram, MA, Vernooij, MW & De Bruijne, M 2016, Asymmetric similarity-weighted ensembles for image segmentation. in 2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings. vol. 2016-June, 7493262, IEEE Computer Society, pp. 273-277, 2016 IEEE 13th International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016, Prague, Czech Republic, 13/04/2016. https://doi.org/10.1109/ISBI.2016.7493262

Asymmetric similarity-weighted ensembles for image segmentation. / Cheplygina, V.; Van Opbroek, A.; Ikram, M.A.; Vernooij, Meike W.; De Bruijne, M.

2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings. Vol. 2016-June IEEE Computer Society, 2016. p. 273-277 7493262.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Asymmetric similarity-weighted ensembles for image segmentation

AU - Cheplygina, V.

AU - Van Opbroek, A.

AU - Ikram, M.A.

AU - Vernooij, Meike W.

AU - De Bruijne, M.

PY - 2016/6/15

Y1 - 2016/6/15

N2 - Supervised classification is widely used for image segmentation. To work effectively, these techniques need large amounts of labeled training data, that is representative of the test data. Different patient groups, different scanners or different scanning protocols can lead to differences between the images, thus representative data might not be available. Transfer learning techniques can be used to account for these differences, thus taking advantage of all the available data acquired with different protocols. We investigate the use of classifier ensembles, where each classifier is weighted according to the similarity between the data it is trained on, and the data it needs to segment. We examine 3 asymmetric similarity measures that can be used in scenarios where no labeled data from a newly introduced scanner or scanning protocol is available. We show that the asymmetry is informative and the direction of measurement needs to be chosen carefully. We also show that a point set similarity measure is robust across different studies, and outperforms state-of-the-art results on a multi-center brain tissue segmentation task.

AB - Supervised classification is widely used for image segmentation. To work effectively, these techniques need large amounts of labeled training data, that is representative of the test data. Different patient groups, different scanners or different scanning protocols can lead to differences between the images, thus representative data might not be available. Transfer learning techniques can be used to account for these differences, thus taking advantage of all the available data acquired with different protocols. We investigate the use of classifier ensembles, where each classifier is weighted according to the similarity between the data it is trained on, and the data it needs to segment. We examine 3 asymmetric similarity measures that can be used in scenarios where no labeled data from a newly introduced scanner or scanning protocol is available. We show that the asymmetry is informative and the direction of measurement needs to be chosen carefully. We also show that a point set similarity measure is robust across different studies, and outperforms state-of-the-art results on a multi-center brain tissue segmentation task.

KW - asymmetry

KW - similarity measure

KW - tissue segmentation

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=84978434492&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2016.7493262

DO - 10.1109/ISBI.2016.7493262

M3 - Conference contribution

VL - 2016-June

SP - 273

EP - 277

BT - 2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings

PB - IEEE Computer Society

ER -

Cheplygina V, Van Opbroek A, Ikram MA, Vernooij MW, De Bruijne M. Asymmetric similarity-weighted ensembles for image segmentation. In 2016 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2016 - Proceedings. Vol. 2016-June. IEEE Computer Society. 2016. p. 273-277. 7493262 https://doi.org/10.1109/ISBI.2016.7493262