Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images

Tong Li; Zhuochen Wang; Yanbo Chen; Lichi Zhang; Yaozong Gao; Feng Shi; Dahong Qian; Qian Wang; Dinggang Shen

doi:10.1007/978-3-030-62469-9_1

Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images

Tong Li, Zhuochen Wang, Yanbo Chen, Lichi Zhang, Yaozong Gao, Feng Shi, Dahong Qian, Qian Wang, Dinggang Shen

Department of Brain and Cognitive Engineering

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

Abstract

Infection segmentation is essential for quantitative assessment in computer-aided management of COVID-19. However, clinical CT images are usually heterogeneous, which are reconstructed by different protocols with varying appearance and voxel spacing due to radiologist preference. Most existing infection segmentation models are only trained using specific types of CT images, which would undermine the performance when applied to other types of CT images. Therefore, it is highly desirable to construct a model that could be applied to heterogeneous CT images in the urgent COVID-19 applications. In this paper, we present a two-stage mapping-segmentation framework for delineating COVID-19 infections from CT images. To compensate for heterogeneity of CT images obtained from different imaging centers, we develop an image-level domain-adaptive process to transform all kinds of images into a target type, and then segment COVID-19 infections accordingly. Experiments show that the infection delineation performance based on our proposed method is superior to the model trained jointly using mixture of all types of images, and is also comparable to those models supervised by using only specific types of images (if applicable).

Original language	English
Title of host publication	Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings
Editors	Jens Petersen, Raúl San José Estépar, Alexander Schmidt-Richberg, Sarah Gerard, Bianca Lassen-Schmidt, Colin Jacobs, Reinhard Beichel, Kensaku Mori
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	3-13
Number of pages	11
ISBN (Print)	9783030624682
DOIs	https://doi.org/10.1007/978-3-030-62469-9_1
Publication status	Published - 2020
Event	2nd International Workshop on Thoracic Image Analysis, TIA 2020 Held in Conjunction with Medical Image Computing and Computer-Assisted Intervention Conference, MICCAI 2020 - Lima, Peru Duration: 2020 Oct 8 → 2020 Oct 8

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12502 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	2nd International Workshop on Thoracic Image Analysis, TIA 2020 Held in Conjunction with Medical Image Computing and Computer-Assisted Intervention Conference, MICCAI 2020
Country	Peru
City	Lima
Period	20/10/8 → 20/10/8

Keywords

COVID-19
Domain-adaption
Image segmentation

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-030-62469-9_1

Fingerprint Dive into the research topics of 'Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images'. Together they form a unique fingerprint.

Cite this

Li, T., Wang, Z., Chen, Y., Zhang, L., Gao, Y., Shi, F., Qian, D., Wang, Q., & Shen, D. (2020). Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images. In J. Petersen, R. San José Estépar, A. Schmidt-Richberg, S. Gerard, B. Lassen-Schmidt, C. Jacobs, R. Beichel, & K. Mori (Eds.), Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings (pp. 3-13). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12502 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-62469-9_1

Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images. / Li, Tong; Wang, Zhuochen; Chen, Yanbo; Zhang, Lichi; Gao, Yaozong; Shi, Feng; Qian, Dahong; Wang, Qian; Shen, Dinggang.

Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings. ed. / Jens Petersen; Raúl San José Estépar; Alexander Schmidt-Richberg; Sarah Gerard; Bianca Lassen-Schmidt; Colin Jacobs; Reinhard Beichel; Kensaku Mori. Springer Science and Business Media Deutschland GmbH, 2020. p. 3-13 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12502 LNCS).

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

Li, T, Wang, Z, Chen, Y, Zhang, L, Gao, Y, Shi, F, Qian, D, Wang, Q & Shen, D 2020, Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images. in J Petersen, R San José Estépar, A Schmidt-Richberg, S Gerard, B Lassen-Schmidt, C Jacobs, R Beichel & K Mori (eds), Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12502 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 3-13, 2nd International Workshop on Thoracic Image Analysis, TIA 2020 Held in Conjunction with Medical Image Computing and Computer-Assisted Intervention Conference, MICCAI 2020, Lima, Peru, 20/10/8. https://doi.org/10.1007/978-3-030-62469-9_1

Li T, Wang Z, Chen Y, Zhang L, Gao Y, Shi F et al. Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images. In Petersen J, San José Estépar R, Schmidt-Richberg A, Gerard S, Lassen-Schmidt B, Jacobs C, Beichel R, Mori K, editors, Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 3-13. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-62469-9_1

Li, Tong ; Wang, Zhuochen ; Chen, Yanbo ; Zhang, Lichi ; Gao, Yaozong ; Shi, Feng ; Qian, Dahong ; Wang, Qian ; Shen, Dinggang. / Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images. Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings. editor / Jens Petersen ; Raúl San José Estépar ; Alexander Schmidt-Richberg ; Sarah Gerard ; Bianca Lassen-Schmidt ; Colin Jacobs ; Reinhard Beichel ; Kensaku Mori. Springer Science and Business Media Deutschland GmbH, 2020. pp. 3-13 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3beed245f34641ee9d4f8e3bcb91f52f,

title = "Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images",

abstract = "Infection segmentation is essential for quantitative assessment in computer-aided management of COVID-19. However, clinical CT images are usually heterogeneous, which are reconstructed by different protocols with varying appearance and voxel spacing due to radiologist preference. Most existing infection segmentation models are only trained using specific types of CT images, which would undermine the performance when applied to other types of CT images. Therefore, it is highly desirable to construct a model that could be applied to heterogeneous CT images in the urgent COVID-19 applications. In this paper, we present a two-stage mapping-segmentation framework for delineating COVID-19 infections from CT images. To compensate for heterogeneity of CT images obtained from different imaging centers, we develop an image-level domain-adaptive process to transform all kinds of images into a target type, and then segment COVID-19 infections accordingly. Experiments show that the infection delineation performance based on our proposed method is superior to the model trained jointly using mixture of all types of images, and is also comparable to those models supervised by using only specific types of images (if applicable).",

keywords = "COVID-19, Domain-adaption, Image segmentation",

author = "Tong Li and Zhuochen Wang and Yanbo Chen and Lichi Zhang and Yaozong Gao and Feng Shi and Dahong Qian and Qian Wang and Dinggang Shen",

note = "Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.; 2nd International Workshop on Thoracic Image Analysis, TIA 2020 Held in Conjunction with Medical Image Computing and Computer-Assisted Intervention Conference, MICCAI 2020 ; Conference date: 08-10-2020 Through 08-10-2020",

year = "2020",

doi = "10.1007/978-3-030-62469-9_1",

language = "English",

isbn = "9783030624682",

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 = "3--13",

editor = "Jens Petersen and {San Jos{\'e} Est{\'e}par}, Ra{\'u}l and Alexander Schmidt-Richberg and Sarah Gerard and Bianca Lassen-Schmidt and Colin Jacobs and Reinhard Beichel and Kensaku Mori",

booktitle = "Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings",

address = "Germany",

}

TY - GEN

T1 - Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images

AU - Li, Tong

AU - Wang, Zhuochen

AU - Chen, Yanbo

AU - Zhang, Lichi

AU - Gao, Yaozong

AU - Shi, Feng

AU - Qian, Dahong

AU - Wang, Qian

AU - Shen, Dinggang

PY - 2020

Y1 - 2020

N2 - Infection segmentation is essential for quantitative assessment in computer-aided management of COVID-19. However, clinical CT images are usually heterogeneous, which are reconstructed by different protocols with varying appearance and voxel spacing due to radiologist preference. Most existing infection segmentation models are only trained using specific types of CT images, which would undermine the performance when applied to other types of CT images. Therefore, it is highly desirable to construct a model that could be applied to heterogeneous CT images in the urgent COVID-19 applications. In this paper, we present a two-stage mapping-segmentation framework for delineating COVID-19 infections from CT images. To compensate for heterogeneity of CT images obtained from different imaging centers, we develop an image-level domain-adaptive process to transform all kinds of images into a target type, and then segment COVID-19 infections accordingly. Experiments show that the infection delineation performance based on our proposed method is superior to the model trained jointly using mixture of all types of images, and is also comparable to those models supervised by using only specific types of images (if applicable).

AB - Infection segmentation is essential for quantitative assessment in computer-aided management of COVID-19. However, clinical CT images are usually heterogeneous, which are reconstructed by different protocols with varying appearance and voxel spacing due to radiologist preference. Most existing infection segmentation models are only trained using specific types of CT images, which would undermine the performance when applied to other types of CT images. Therefore, it is highly desirable to construct a model that could be applied to heterogeneous CT images in the urgent COVID-19 applications. In this paper, we present a two-stage mapping-segmentation framework for delineating COVID-19 infections from CT images. To compensate for heterogeneity of CT images obtained from different imaging centers, we develop an image-level domain-adaptive process to transform all kinds of images into a target type, and then segment COVID-19 infections accordingly. Experiments show that the infection delineation performance based on our proposed method is superior to the model trained jointly using mixture of all types of images, and is also comparable to those models supervised by using only specific types of images (if applicable).

KW - COVID-19

KW - Domain-adaption

KW - Image segmentation

UR - http://www.scopus.com/inward/record.url?scp=85097257561&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85097257561&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-62469-9_1

DO - 10.1007/978-3-030-62469-9_1

M3 - Conference contribution

AN - SCOPUS:85097257561

SN - 9783030624682

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

SP - 3

EP - 13

BT - Thoracic Image Analysis - Second International Workshop, TIA 2020, Held in Conjunction with MICCAI 2020, Proceedings

A2 - Petersen, Jens

A2 - San José Estépar, Raúl

A2 - Schmidt-Richberg, Alexander

A2 - Gerard, Sarah

A2 - Lassen-Schmidt, Bianca

A2 - Jacobs, Colin

A2 - Beichel, Reinhard

A2 - Mori, Kensaku

PB - Springer Science and Business Media Deutschland GmbH

T2 - 2nd International Workshop on Thoracic Image Analysis, TIA 2020 Held in Conjunction with Medical Image Computing and Computer-Assisted Intervention Conference, MICCAI 2020

Y2 - 8 October 2020 through 8 October 2020

ER -