Weakly Supervised Volumetric Image Segmentation with Deformed Templates

Udaranga Wickramasinghe; Patrick Jensen; Mian Shah; Jiancheng Yang; Pascal Fua

doi:10.1007/978-3-031-16443-9_41

Weakly Supervised Volumetric Image Segmentation with Deformed Templates

Udaranga Wickramasinghe^*, Patrick Jensen, Mian Shah, Jiancheng Yang, Pascal Fua

^*Corresponding author for this work

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

35 Downloads (Orbit)

Abstract

There are many approaches to weakly-supervised training of networks to segment 2D images. By contrast, existing approaches to segmenting volumetric images rely on full-supervision of a subset of 2D slices of the 3D volume. We propose an approach to volume segmentation that is truly weakly-supervised in the sense that we only need to provide a sparse set of 3D points on the surface of target objects instead of detailed 2D masks. We use the 3D points to deform a 3D template so that it roughly matches the target object outlines and we introduce an architecture that exploits the supervision it provides to train a network to find accurate boundaries. We evaluate our approach on Computed Tomography (CT), Magnetic Resonance Imagery (MRI) and Electron Microscopy (EM) image datasets and show that it substantially reduces the required amount of effort.

Original language	English
Title of host publication	Proceedings of the 25th International Conference of Medical Image Computing and Computer Assisted Intervention, MICCAI 2022
Volume	13435
Publisher	Springer
Publication date	2022
Pages	422-432
ISBN (Print)	978-3-031-16442-2
ISBN (Electronic)	978-3-031-16443-9
DOIs	https://doi.org/10.1007/978-3-031-16443-9_41
Publication status	Published - 2022
Event	25th International Conference on Medical Image Computing and Computer-Assisted Intervention - Resorts World Convention Centre Singapore, Singapore, Singapore Duration: 18 Sept 2022 → 22 Sept 2022 Conference number: 25 https://conferences.miccai.org/2022/en/

Conference

Conference	25th International Conference on Medical Image Computing and Computer-Assisted Intervention
Number	25
Location	Resorts World Convention Centre Singapore
Country/Territory	Singapore
City	Singapore
Period	18/09/2022 → 22/09/2022
Internet address	https://conferences.miccai.org/2022/en/

Series	Lecture Notes in Computer Science
Volume	13435
ISSN	0302-9743

Access to Document

10.1007/978-3-031-16443-9_41

FulltextAccepted author manuscript, 1.37 MB

OpenUrl availability

Full text

Cite this

@inproceedings{3d21270bfd5145b49a52571a15117204,

title = "Weakly Supervised Volumetric Image Segmentation with Deformed Templates",

abstract = "There are many approaches to weakly-supervised training of networks to segment 2D images. By contrast, existing approaches to segmenting volumetric images rely on full-supervision of a subset of 2D slices of the 3D volume. We propose an approach to volume segmentation that is truly weakly-supervised in the sense that we only need to provide a sparse set of 3D points on the surface of target objects instead of detailed 2D masks. We use the 3D points to deform a 3D template so that it roughly matches the target object outlines and we introduce an architecture that exploits the supervision it provides to train a network to find accurate boundaries. We evaluate our approach on Computed Tomography (CT), Magnetic Resonance Imagery (MRI) and Electron Microscopy (EM) image datasets and show that it substantially reduces the required amount of effort.",

author = "Udaranga Wickramasinghe and Patrick Jensen and Mian Shah and Jiancheng Yang and Pascal Fua",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 ; Conference date: 18-09-2022 Through 22-09-2022",

year = "2022",

doi = "10.1007/978-3-031-16443-9_41",

language = "English",

isbn = " 978-3-031-16442-2",

volume = "13435",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "422--432",

booktitle = "Proceedings of the 25th International Conference of Medical Image Computing and Computer Assisted Intervention, MICCAI 2022",

url = "https://conferences.miccai.org/2022/en/",

}

Wickramasinghe, U, Jensen, P, Shah, M, Yang, J & Fua, P 2022, Weakly Supervised Volumetric Image Segmentation with Deformed Templates. in Proceedings of the 25th International Conference of Medical Image Computing and Computer Assisted Intervention, MICCAI 2022. vol. 13435, Springer, Lecture Notes in Computer Science, vol. 13435, pp. 422-432, 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, Singapore, Singapore, 18/09/2022. https://doi.org/10.1007/978-3-031-16443-9_41

Weakly Supervised Volumetric Image Segmentation with Deformed Templates. / Wickramasinghe, Udaranga; Jensen, Patrick; Shah, Mian et al.
Proceedings of the 25th International Conference of Medical Image Computing and Computer Assisted Intervention, MICCAI 2022. Vol. 13435 Springer, 2022. p. 422-432 (Lecture Notes in Computer Science, Vol. 13435).

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

TY - GEN

T1 - Weakly Supervised Volumetric Image Segmentation with Deformed Templates

AU - Wickramasinghe, Udaranga

AU - Jensen, Patrick

AU - Shah, Mian

AU - Yang, Jiancheng

AU - Fua, Pascal

N1 - Conference code: 25

PY - 2022

Y1 - 2022

N2 - There are many approaches to weakly-supervised training of networks to segment 2D images. By contrast, existing approaches to segmenting volumetric images rely on full-supervision of a subset of 2D slices of the 3D volume. We propose an approach to volume segmentation that is truly weakly-supervised in the sense that we only need to provide a sparse set of 3D points on the surface of target objects instead of detailed 2D masks. We use the 3D points to deform a 3D template so that it roughly matches the target object outlines and we introduce an architecture that exploits the supervision it provides to train a network to find accurate boundaries. We evaluate our approach on Computed Tomography (CT), Magnetic Resonance Imagery (MRI) and Electron Microscopy (EM) image datasets and show that it substantially reduces the required amount of effort.

AB - There are many approaches to weakly-supervised training of networks to segment 2D images. By contrast, existing approaches to segmenting volumetric images rely on full-supervision of a subset of 2D slices of the 3D volume. We propose an approach to volume segmentation that is truly weakly-supervised in the sense that we only need to provide a sparse set of 3D points on the surface of target objects instead of detailed 2D masks. We use the 3D points to deform a 3D template so that it roughly matches the target object outlines and we introduce an architecture that exploits the supervision it provides to train a network to find accurate boundaries. We evaluate our approach on Computed Tomography (CT), Magnetic Resonance Imagery (MRI) and Electron Microscopy (EM) image datasets and show that it substantially reduces the required amount of effort.

U2 - 10.1007/978-3-031-16443-9_41

DO - 10.1007/978-3-031-16443-9_41

M3 - Article in proceedings

AN - SCOPUS:85138994839

SN - 978-3-031-16442-2

VL - 13435

T3 - Lecture Notes in Computer Science

SP - 422

EP - 432

BT - Proceedings of the 25th International Conference of Medical Image Computing and Computer Assisted Intervention, MICCAI 2022

PB - Springer

T2 - 25th International Conference on Medical Image Computing and Computer-Assisted Intervention

Y2 - 18 September 2022 through 22 September 2022

ER -

Wickramasinghe U, Jensen P, Shah M, Yang J, Fua P. Weakly Supervised Volumetric Image Segmentation with Deformed Templates. In Proceedings of the 25th International Conference of Medical Image Computing and Computer Assisted Intervention, MICCAI 2022. Vol. 13435. Springer. 2022. p. 422-432. (Lecture Notes in Computer Science, Vol. 13435). doi: 10.1007/978-3-031-16443-9_41

Weakly Supervised Volumetric Image Segmentation with Deformed Templates

Abstract

Conference

Access to Document

OpenUrl availability

Fingerprint

Cite this