Fetal Cortex Segmentation with Topology and Thickness Loss Constraints

Liu Li; Qiang Ma; Zeju Li; Cheng Ouyang; Weitong Zhang; Anthony Price; Vanessa Kyriakopoulou; Lucilio C. Grande; Antonis Makropoulos; Joseph Hajnal; Daniel Rueckert; Bernhard Kainz; Amir Alansary

doi:10.1007/978-3-031-23223-7_11

Fetal Cortex Segmentation with Topology and Thickness Loss Constraints

Liu Li^*, Qiang Ma, Zeju Li, Cheng Ouyang, Weitong Zhang, Anthony Price, Vanessa Kyriakopoulou, Lucilio C. Grande, Antonis Makropoulos, Joseph Hajnal, Daniel Rueckert, Bernhard Kainz, Amir Alansary

^*Corresponding author for this work

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

3 Citations (Scopus)

Abstract

The segmentation of the fetal cerebral cortex from magnetic resonance imaging (MRI) is an important tool for neurobiological research about the developing human brain. Manual segmentation is difficult and time-consuming. Limited image resolution and partial volume effects introduce errors and labeling noise when attempting to automate the process through machine learning. The significant morphological changes observed during brain growth pose additional challenges for learning-based image segmentation methods, which may drastically increase the amount of necessary training data. In this paper, we propose a framework to learn from noisy labels by using additional regularization via shape priors for the accurate segmentation of the cortical gray matter (CGM) in 3D. Firstly, we introduce a novel structure consistency loss based on persistent homology analysis of the cortical topology. Secondly, a regularization loss term is proposed by integrating assumptions about the cortical thickness within each sample. Our experiments on the developing human connectome project (dHCP) dataset show that our method can predict accurate CGM segmentation learned from noisy labels.

Original language	English
Title of host publication	Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, Held in Conjunction with MICCAI 2022, Proceedings
Editors	John S.H. Baxter, Islem Rekik, Roy Eagleson, Luping Zhou, Tanveer Syeda-Mahmood, Hongzhi Wang, Mustafa Hajij
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	123-133
Number of pages	11
ISBN (Print)	9783031232220
DOIs	https://doi.org/10.1007/978-3-031-23223-7_11
Publication status	Published - 2022
Event	1st International Workshop on Ethical and Philosophical Issues in Medical Imaging, EPIMI 2022, the 12th International Workshop on Multimodal Learning and Fusion Across Scales for Clinical Decision Support, ML-CDS 2022, and the 2nd International Workshop on Topological Data Analysis for Biomedical Imaging, TDA4BiomedicalImaging 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 - Singapore, Singapore Duration: 18 Sept 2022 → 22 Sept 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13755 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	1st International Workshop on Ethical and Philosophical Issues in Medical Imaging, EPIMI 2022, the 12th International Workshop on Multimodal Learning and Fusion Across Scales for Clinical Decision Support, ML-CDS 2022, and the 2nd International Workshop on Topological Data Analysis for Biomedical Imaging, TDA4BiomedicalImaging 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	18/09/2022 → 22/09/2022

Access to Document

10.1007/978-3-031-23223-7_11

Cite this

Li, L., Ma, Q., Li, Z., Ouyang, C., Zhang, W., Price, A., Kyriakopoulou, V., Grande, L. C., Makropoulos, A., Hajnal, J., Rueckert, D., Kainz, B., & Alansary, A. (2022). Fetal Cortex Segmentation with Topology and Thickness Loss Constraints. In J. S. H. Baxter, I. Rekik, R. Eagleson, L. Zhou, T. Syeda-Mahmood, H. Wang, & M. Hajij (Eds.), Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, Held in Conjunction with MICCAI 2022, Proceedings (pp. 123-133). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13755 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-23223-7_11

Li, Liu ; Ma, Qiang ; Li, Zeju et al. / Fetal Cortex Segmentation with Topology and Thickness Loss Constraints. Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, Held in Conjunction with MICCAI 2022, Proceedings. editor / John S.H. Baxter ; Islem Rekik ; Roy Eagleson ; Luping Zhou ; Tanveer Syeda-Mahmood ; Hongzhi Wang ; Mustafa Hajij. Springer Science and Business Media Deutschland GmbH, 2022. pp. 123-133 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inbook{7165e861863e4b8797e641099ce3b7e2,

title = "Fetal Cortex Segmentation with Topology and Thickness Loss Constraints",

abstract = "The segmentation of the fetal cerebral cortex from magnetic resonance imaging (MRI) is an important tool for neurobiological research about the developing human brain. Manual segmentation is difficult and time-consuming. Limited image resolution and partial volume effects introduce errors and labeling noise when attempting to automate the process through machine learning. The significant morphological changes observed during brain growth pose additional challenges for learning-based image segmentation methods, which may drastically increase the amount of necessary training data. In this paper, we propose a framework to learn from noisy labels by using additional regularization via shape priors for the accurate segmentation of the cortical gray matter (CGM) in 3D. Firstly, we introduce a novel structure consistency loss based on persistent homology analysis of the cortical topology. Secondly, a regularization loss term is proposed by integrating assumptions about the cortical thickness within each sample. Our experiments on the developing human connectome project (dHCP) dataset show that our method can predict accurate CGM segmentation learned from noisy labels.",

author = "Liu Li and Qiang Ma and Zeju Li and Cheng Ouyang and Weitong Zhang and Anthony Price and Vanessa Kyriakopoulou and Grande, {Lucilio C.} and Antonis Makropoulos and Joseph Hajnal and Daniel Rueckert and Bernhard Kainz and Amir Alansary",

note = "Funding Information: Data in this work were provided by ERC Grant Agreement no. [319456]. We are grateful to the families who generously supported this trial. Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 1st International Workshop on Ethical and Philosophical Issues in Medical Imaging, EPIMI 2022, the 12th International Workshop on Multimodal Learning and Fusion Across Scales for Clinical Decision Support, ML-CDS 2022, and the 2nd International Workshop on Topological Data Analysis for Biomedical Imaging, TDA4BiomedicalImaging 2022, held in conjunction with the 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 = "10.1007/978-3-031-23223-7_11",

language = "English",

isbn = "9783031232220",

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 = "123--133",

editor = "Baxter, {John S.H.} and Islem Rekik and Roy Eagleson and Luping Zhou and Tanveer Syeda-Mahmood and Hongzhi Wang and Mustafa Hajij",

booktitle = "Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, Held in Conjunction with MICCAI 2022, Proceedings",

address = "Germany",

}

Li, L, Ma, Q, Li, Z, Ouyang, C, Zhang, W, Price, A , Kyriakopoulou, V, Grande, LC, Makropoulos, A, Hajnal, J, Rueckert, D, Kainz, B & Alansary, A 2022, Fetal Cortex Segmentation with Topology and Thickness Loss Constraints. in JSH Baxter, I Rekik, R Eagleson, L Zhou, T Syeda-Mahmood, H Wang & M Hajij (eds), Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, 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. 13755 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 123-133, 1st International Workshop on Ethical and Philosophical Issues in Medical Imaging, EPIMI 2022, the 12th International Workshop on Multimodal Learning and Fusion Across Scales for Clinical Decision Support, ML-CDS 2022, and the 2nd International Workshop on Topological Data Analysis for Biomedical Imaging, TDA4BiomedicalImaging 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022, Singapore, Singapore, 18/09/2022. https://doi.org/10.1007/978-3-031-23223-7_11

Fetal Cortex Segmentation with Topology and Thickness Loss Constraints. / Li, Liu; Ma, Qiang; Li, Zeju et al.
Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, Held in Conjunction with MICCAI 2022, Proceedings. ed. / John S.H. Baxter; Islem Rekik; Roy Eagleson; Luping Zhou; Tanveer Syeda-Mahmood; Hongzhi Wang; Mustafa Hajij. Springer Science and Business Media Deutschland GmbH, 2022. p. 123-133 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13755 LNCS).

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

TY - CHAP

T1 - Fetal Cortex Segmentation with Topology and Thickness Loss Constraints

AU - Li, Liu

AU - Ma, Qiang

AU - Li, Zeju

AU - Ouyang, Cheng

AU - Zhang, Weitong

AU - Price, Anthony

AU - Kyriakopoulou, Vanessa

AU - Grande, Lucilio C.

AU - Makropoulos, Antonis

AU - Hajnal, Joseph

AU - Rueckert, Daniel

AU - Kainz, Bernhard

AU - Alansary, Amir

N1 - Funding Information: Data in this work were provided by ERC Grant Agreement no. [319456]. We are grateful to the families who generously supported this trial. Publisher Copyright: © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

PY - 2022

Y1 - 2022

N2 - The segmentation of the fetal cerebral cortex from magnetic resonance imaging (MRI) is an important tool for neurobiological research about the developing human brain. Manual segmentation is difficult and time-consuming. Limited image resolution and partial volume effects introduce errors and labeling noise when attempting to automate the process through machine learning. The significant morphological changes observed during brain growth pose additional challenges for learning-based image segmentation methods, which may drastically increase the amount of necessary training data. In this paper, we propose a framework to learn from noisy labels by using additional regularization via shape priors for the accurate segmentation of the cortical gray matter (CGM) in 3D. Firstly, we introduce a novel structure consistency loss based on persistent homology analysis of the cortical topology. Secondly, a regularization loss term is proposed by integrating assumptions about the cortical thickness within each sample. Our experiments on the developing human connectome project (dHCP) dataset show that our method can predict accurate CGM segmentation learned from noisy labels.

AB - The segmentation of the fetal cerebral cortex from magnetic resonance imaging (MRI) is an important tool for neurobiological research about the developing human brain. Manual segmentation is difficult and time-consuming. Limited image resolution and partial volume effects introduce errors and labeling noise when attempting to automate the process through machine learning. The significant morphological changes observed during brain growth pose additional challenges for learning-based image segmentation methods, which may drastically increase the amount of necessary training data. In this paper, we propose a framework to learn from noisy labels by using additional regularization via shape priors for the accurate segmentation of the cortical gray matter (CGM) in 3D. Firstly, we introduce a novel structure consistency loss based on persistent homology analysis of the cortical topology. Secondly, a regularization loss term is proposed by integrating assumptions about the cortical thickness within each sample. Our experiments on the developing human connectome project (dHCP) dataset show that our method can predict accurate CGM segmentation learned from noisy labels.

UR - http://www.scopus.com/inward/record.url?scp=85144814906&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-23223-7_11

DO - 10.1007/978-3-031-23223-7_11

M3 - Conference paper

AN - SCOPUS:85144814906

SN - 9783031232220

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

SP - 123

EP - 133

BT - Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, Held in Conjunction with MICCAI 2022, Proceedings

A2 - Baxter, John S.H.

A2 - Rekik, Islem

A2 - Eagleson, Roy

A2 - Zhou, Luping

A2 - Syeda-Mahmood, Tanveer

A2 - Wang, Hongzhi

A2 - Hajij, Mustafa

PB - Springer Science and Business Media Deutschland GmbH

T2 - 1st International Workshop on Ethical and Philosophical Issues in Medical Imaging, EPIMI 2022, the 12th International Workshop on Multimodal Learning and Fusion Across Scales for Clinical Decision Support, ML-CDS 2022, and the 2nd International Workshop on Topological Data Analysis for Biomedical Imaging, TDA4BiomedicalImaging 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022

Y2 - 18 September 2022 through 22 September 2022

ER -

Li L, Ma Q, Li Z, Ouyang C, Zhang W, Price A et al. Fetal Cortex Segmentation with Topology and Thickness Loss Constraints. In Baxter JSH, Rekik I, Eagleson R, Zhou L, Syeda-Mahmood T, Wang H, Hajij M, editors, Ethical and Philosophical Issues in Medical Imaging, Multimodal Learning and Fusion Across Scales for Clinical Decision Support, and Topological Data Analysis for Biomedical Imaging - 1st International Workshop, EPIMI 2022, 12th International Workshop, ML-CDS 2022, 2nd International Workshop, TDA4BiomedicalImaging, Held in Conjunction with MICCAI 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 123-133. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-23223-7_11

Fetal Cortex Segmentation with Topology and Thickness Loss Constraints

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this