Unsupervised Site Adaptation by Intra-site Variability Alignment

Shaya Goodman; Shira Kasten Serlin; Hayit Greenspan; Jacob Goldberger

doi:10.1007/978-3-031-16852-9_6

Unsupervised Site Adaptation by Intra-site Variability Alignment

Shaya Goodman, Shira Kasten Serlin, Hayit Greenspan, Jacob Goldberger

Bar-Ilan University - The Alexander Kofkin Faculty of Engineering

Tel Aviv University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

A medical imaging network that was trained on a particular source domain usually suffers significant performance degradation when transferred to a different target domain. This is known as the domain-shift problem. In this study, we propose a general method for transfer knowledge from a source site with labeled data to a target site where only unlabeled data is available. We leverage the variability that is often present within each site, the intra-site variability, and propose an unsupervised site adaptation method that jointly aligns the intra-site data variability in the source and target sites while training the network on the labeled source site data. We applied our method to several medical MRI image segmentation tasks and show that it consistently outperforms state-of-the-art methods.

Original language	English
Title of host publication	Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings
Editors	Konstantinos Kamnitsas, Lisa Koch, Mobarakol Islam, Ziyue Xu, Jorge Cardoso, Qi Dou, Nicola Rieke, Sotirios Tsaftaris
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	56-65
Number of pages	10
ISBN (Print)	9783031168512
DOIs	https://doi.org/10.1007/978-3-031-16852-9_6
State	Published - 2022
Event	4th MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 - Singapore, Singapore Duration: 22 Sep 2022 → 22 Sep 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13542 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	4th MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
Country/Territory	Singapore
City	Singapore
Period	22/09/22 → 22/09/22

Bibliographical note

Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Keywords

Intra-site variability
MRI segmentation
UDA
Unsupervised domain adaptation

Access to Document

10.1007/978-3-031-16852-9_6

Cite this

Goodman, S., Kasten Serlin, S., Greenspan, H., & Goldberger, J. (2022). Unsupervised Site Adaptation by Intra-site Variability Alignment. In K. Kamnitsas, L. Koch, M. Islam, Z. Xu, J. Cardoso, Q. Dou, N. Rieke, & S. Tsaftaris (Eds.), Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings (pp. 56-65). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13542 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-16852-9_6

Goodman, Shaya ; Kasten Serlin, Shira ; Greenspan, Hayit et al. / Unsupervised Site Adaptation by Intra-site Variability Alignment. Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings. editor / Konstantinos Kamnitsas ; Lisa Koch ; Mobarakol Islam ; Ziyue Xu ; Jorge Cardoso ; Qi Dou ; Nicola Rieke ; Sotirios Tsaftaris. Springer Science and Business Media Deutschland GmbH, 2022. pp. 56-65 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{f78cd8cbd593441fb37b1f5f9eed9799,

title = "Unsupervised Site Adaptation by Intra-site Variability Alignment",

abstract = "A medical imaging network that was trained on a particular source domain usually suffers significant performance degradation when transferred to a different target domain. This is known as the domain-shift problem. In this study, we propose a general method for transfer knowledge from a source site with labeled data to a target site where only unlabeled data is available. We leverage the variability that is often present within each site, the intra-site variability, and propose an unsupervised site adaptation method that jointly aligns the intra-site data variability in the source and target sites while training the network on the labeled source site data. We applied our method to several medical MRI image segmentation tasks and show that it consistently outperforms state-of-the-art methods.",

keywords = "Intra-site variability, MRI segmentation, UDA, Unsupervised domain adaptation",

author = "Shaya Goodman and {Kasten Serlin}, Shira and Hayit Greenspan and Jacob Goldberger",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 4th MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 ; Conference date: 22-09-2022 Through 22-09-2022",

year = "2022",

doi = "10.1007/978-3-031-16852-9_6",

language = "אנגלית",

isbn = "9783031168512",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "56--65",

editor = "Konstantinos Kamnitsas and Lisa Koch and Mobarakol Islam and Ziyue Xu and Jorge Cardoso and Qi Dou and Nicola Rieke and Sotirios Tsaftaris",

booktitle = "Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings",

address = "גרמניה",

}

Goodman, S, Kasten Serlin, S, Greenspan, H & Goldberger, J 2022, Unsupervised Site Adaptation by Intra-site Variability Alignment. in K Kamnitsas, L Koch, M Islam, Z Xu, J Cardoso, Q Dou, N Rieke & S Tsaftaris (eds), Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13542 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 56-65, 4th MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022, Singapore, Singapore, 22/09/22. https://doi.org/10.1007/978-3-031-16852-9_6

Unsupervised Site Adaptation by Intra-site Variability Alignment. / Goodman, Shaya; Kasten Serlin, Shira; Greenspan, Hayit et al.
Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings. ed. / Konstantinos Kamnitsas; Lisa Koch; Mobarakol Islam; Ziyue Xu; Jorge Cardoso; Qi Dou; Nicola Rieke; Sotirios Tsaftaris. Springer Science and Business Media Deutschland GmbH, 2022. p. 56-65 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13542 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Unsupervised Site Adaptation by Intra-site Variability Alignment

AU - Goodman, Shaya

AU - Kasten Serlin, Shira

AU - Greenspan, Hayit

AU - Goldberger, Jacob

PY - 2022

Y1 - 2022

N2 - A medical imaging network that was trained on a particular source domain usually suffers significant performance degradation when transferred to a different target domain. This is known as the domain-shift problem. In this study, we propose a general method for transfer knowledge from a source site with labeled data to a target site where only unlabeled data is available. We leverage the variability that is often present within each site, the intra-site variability, and propose an unsupervised site adaptation method that jointly aligns the intra-site data variability in the source and target sites while training the network on the labeled source site data. We applied our method to several medical MRI image segmentation tasks and show that it consistently outperforms state-of-the-art methods.

AB - A medical imaging network that was trained on a particular source domain usually suffers significant performance degradation when transferred to a different target domain. This is known as the domain-shift problem. In this study, we propose a general method for transfer knowledge from a source site with labeled data to a target site where only unlabeled data is available. We leverage the variability that is often present within each site, the intra-site variability, and propose an unsupervised site adaptation method that jointly aligns the intra-site data variability in the source and target sites while training the network on the labeled source site data. We applied our method to several medical MRI image segmentation tasks and show that it consistently outperforms state-of-the-art methods.

KW - Intra-site variability

KW - MRI segmentation

KW - UDA

KW - Unsupervised domain adaptation

UR - http://www.scopus.com/inward/record.url?scp=85140472370&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-16852-9_6

DO - 10.1007/978-3-031-16852-9_6

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85140472370

SN - 9783031168512

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

SP - 56

EP - 65

BT - Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings

A2 - Kamnitsas, Konstantinos

A2 - Koch, Lisa

A2 - Islam, Mobarakol

A2 - Xu, Ziyue

A2 - Cardoso, Jorge

A2 - Dou, Qi

A2 - Rieke, Nicola

A2 - Tsaftaris, Sotirios

PB - Springer Science and Business Media Deutschland GmbH

T2 - 4th MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022

Y2 - 22 September 2022 through 22 September 2022

ER -

Goodman S, Kasten Serlin S, Greenspan H, Goldberger J. Unsupervised Site Adaptation by Intra-site Variability Alignment. In Kamnitsas K, Koch L, Islam M, Xu Z, Cardoso J, Dou Q, Rieke N, Tsaftaris S, editors, Domain Adaptation and Representation Transfer - 4th MICCAI Workshop, DART 2022, Held in Conjunction with MICCAI 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 56-65. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-16852-9_6