Fluid Segmentation in OCT with an Improved Convolutional Neural Network

Gang Xing; Jianqin Lei; Xiayu Xu

doi:10.1145/3469678.3469699

Fluid Segmentation in OCT with an Improved Convolutional Neural Network

Gang Xing
, Jianqin Lei
, Xiayu Xu

Xi'an Jiaotong University
The First Affiliated Hospital of Xi’an Jiaotong University

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

Abstract

Multi-scale pathological fluid segmentation is of great importance for the diagnosis and treatment of various eye diseases such as neovascular age-related macular degeneration (nAMD) and diabetic macular edema (DME). Despite significant progress in recent years, there are still several important issues remain unsolved. First, abnormal fluid lesions in optical coherence tomography (OCT) show large variations in location, size, and shape. Second, fluid lesions are contiguous regions with smooth surfaces and without holes inside. In this study, we introduce an adapted fully convolutional neural network (FCN) architecture to improve the ability of the network to extract multi-scale fluid lesions in OCT. Then we introduced a novel curvature loss term to regularize the shape prior in the loss function. The proposed method was extensively evaluated on RETOUCH dataset with a mean Dice score (DSC) of 0.767 and mean absolute volume difference (AVD) of 0.036 mm3, which improved significantly compared with the state-of-the-art methods.

Original language	English
Title of host publication	Proceedings of the 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021
Editors	Bin Chen
Publisher	Association for Computing Machinery
ISBN (Electronic)	9781450389297
DOIs	https://doi.org/10.1145/3469678.3469699
State	Published - 20 Jul 2021
Externally published	Yes
Event	5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021 - Hangzhou, China Duration: 20 Jul 2021 → 21 Jul 2021

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021
Country/Territory	China
City	Hangzhou
Period	20/07/21 → 21/07/21

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

Image segmentation
Loss function
Optical coherence tomography
Pathological fluid

Access to Document

10.1145/3469678.3469699

Cite this

Xing, G., Lei, J., & Xu, X. (2021). Fluid Segmentation in OCT with an Improved Convolutional Neural Network. In B. Chen (Ed.), Proceedings of the 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021 Article 3469699 (ACM International Conference Proceeding Series). Association for Computing Machinery. https://doi.org/10.1145/3469678.3469699

@inproceedings{c2059d6346a64d70ba09fa94ca8fb5a0,

title = "Fluid Segmentation in OCT with an Improved Convolutional Neural Network",

abstract = "Multi-scale pathological fluid segmentation is of great importance for the diagnosis and treatment of various eye diseases such as neovascular age-related macular degeneration (nAMD) and diabetic macular edema (DME). Despite significant progress in recent years, there are still several important issues remain unsolved. First, abnormal fluid lesions in optical coherence tomography (OCT) show large variations in location, size, and shape. Second, fluid lesions are contiguous regions with smooth surfaces and without holes inside. In this study, we introduce an adapted fully convolutional neural network (FCN) architecture to improve the ability of the network to extract multi-scale fluid lesions in OCT. Then we introduced a novel curvature loss term to regularize the shape prior in the loss function. The proposed method was extensively evaluated on RETOUCH dataset with a mean Dice score (DSC) of 0.767 and mean absolute volume difference (AVD) of 0.036 mm3, which improved significantly compared with the state-of-the-art methods.",

keywords = "Image segmentation, Loss function, Optical coherence tomography, Pathological fluid",

author = "Gang Xing and Jianqin Lei and Xiayu Xu",

note = "Publisher Copyright: {\textcopyright} 2021 ACM.; 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021 ; Conference date: 20-07-2021 Through 21-07-2021",

year = "2021",

month = jul,

day = "20",

doi = "10.1145/3469678.3469699",

language = "英语",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

editor = "Bin Chen",

booktitle = "Proceedings of the 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021",

}

Xing, G, Lei, J & Xu, X 2021, Fluid Segmentation in OCT with an Improved Convolutional Neural Network. in B Chen (ed.), Proceedings of the 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021., 3469699, ACM International Conference Proceeding Series, Association for Computing Machinery, 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021, Hangzhou, China, 20/07/21. https://doi.org/10.1145/3469678.3469699

Fluid Segmentation in OCT with an Improved Convolutional Neural Network. / Xing, Gang; Lei, Jianqin; Xu, Xiayu.
Proceedings of the 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021. ed. / Bin Chen. Association for Computing Machinery, 2021. 3469699 (ACM International Conference Proceeding Series).

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

TY - GEN

T1 - Fluid Segmentation in OCT with an Improved Convolutional Neural Network

AU - Xing, Gang

AU - Lei, Jianqin

AU - Xu, Xiayu

PY - 2021/7/20

Y1 - 2021/7/20

N2 - Multi-scale pathological fluid segmentation is of great importance for the diagnosis and treatment of various eye diseases such as neovascular age-related macular degeneration (nAMD) and diabetic macular edema (DME). Despite significant progress in recent years, there are still several important issues remain unsolved. First, abnormal fluid lesions in optical coherence tomography (OCT) show large variations in location, size, and shape. Second, fluid lesions are contiguous regions with smooth surfaces and without holes inside. In this study, we introduce an adapted fully convolutional neural network (FCN) architecture to improve the ability of the network to extract multi-scale fluid lesions in OCT. Then we introduced a novel curvature loss term to regularize the shape prior in the loss function. The proposed method was extensively evaluated on RETOUCH dataset with a mean Dice score (DSC) of 0.767 and mean absolute volume difference (AVD) of 0.036 mm3, which improved significantly compared with the state-of-the-art methods.

AB - Multi-scale pathological fluid segmentation is of great importance for the diagnosis and treatment of various eye diseases such as neovascular age-related macular degeneration (nAMD) and diabetic macular edema (DME). Despite significant progress in recent years, there are still several important issues remain unsolved. First, abnormal fluid lesions in optical coherence tomography (OCT) show large variations in location, size, and shape. Second, fluid lesions are contiguous regions with smooth surfaces and without holes inside. In this study, we introduce an adapted fully convolutional neural network (FCN) architecture to improve the ability of the network to extract multi-scale fluid lesions in OCT. Then we introduced a novel curvature loss term to regularize the shape prior in the loss function. The proposed method was extensively evaluated on RETOUCH dataset with a mean Dice score (DSC) of 0.767 and mean absolute volume difference (AVD) of 0.036 mm3, which improved significantly compared with the state-of-the-art methods.

KW - Image segmentation

KW - Loss function

KW - Optical coherence tomography

KW - Pathological fluid

UR - https://www.scopus.com/pages/publications/85120525434

U2 - 10.1145/3469678.3469699

DO - 10.1145/3469678.3469699

M3 - 会议稿件

AN - SCOPUS:85120525434

T3 - ACM International Conference Proceeding Series

BT - Proceedings of the 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021

A2 - Chen, Bin

PB - Association for Computing Machinery

T2 - 5th International Conference on Biological Information and Biomedical Engineering, BIBE 2021

Y2 - 20 July 2021 through 21 July 2021

ER -

Fluid Segmentation in OCT with an Improved Convolutional Neural Network

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this