Deep reinforcement active learning for medical image classification

Jingwen Wang; Yuguang Yan; Yubing Zhang; Guiping Cao; Ming Yang; Michael K. Ng

doi:10.1007/978-3-030-59710-8_4

Deep reinforcement active learning for medical image classification

Jingwen Wang, Yuguang Yan, Yubing Zhang^*, Guiping Cao, Ming Yang, Michael K. Ng

^*Corresponding author for this work

Department of Mathematics

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

16 Citations (Scopus)

Abstract

In this paper, we propose a deep reinforcement learning algorithm for active learning on medical image data. Although deep learning has achieved great success on medical image processing, it relies on a large number of labeled data for training, which is expensive and time-consuming. Active learning, which follows a strategy to select and annotate informative samples, is an effective approach to alleviate this issue. However, most existing methods of active learning adopt a hand-design strategy, which cannot handle the dynamic procedure of classifier training. To address this issue, we model the procedure of active learning as a Markov decision process, and propose a deep reinforcement learning algorithm to learn a dynamic policy for active learning. To achieve this, we employ the actor-critic approach, and apply the deep deterministic policy gradient algorithm to train the model. We conduct experiments on two kinds of medical image data sets, and the results demonstrate that our method is able to learn better strategy compared with the existing hand-design ones.

Original language	English
Title of host publication	Medical Image Computing and Computer Assisted Intervention – MICCAI 2020
Subtitle of host publication	23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part I
Editors	Anne L. Martel, Purang Abolmaesumi, Danail Stoyanov, Diana Mateus, Maria A. Zuluaga, S. Kevin Zhou, Daniel Racoceanu, Leo Joskowicz
Publisher	Springer Cham
Pages	33-42
Number of pages	10
Edition	1st
ISBN (Electronic)	9783030597108
ISBN (Print)	9783030597092
DOIs	https://doi.org/10.1007/978-3-030-59710-8_4
Publication status	Published - 29 Sept 2020
Event	23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020 - Lima, Peru Duration: 4 Oct 2020 → 8 Oct 2020 https://link.springer.com/book/10.1007/978-3-030-59710-8

Publication series

Name	Lecture Notes in Computer Science
Volume	12261
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Name	Image Processing, Computer Vision, Pattern Recognition, and Graphics

Name	MICCAI: International Conference on Medical Image Computing and Computer-Assisted Intervention

Conference

Conference	23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020
Country/Territory	Peru
City	Lima
Period	4/10/20 → 8/10/20
Internet address	https://link.springer.com/book/10.1007/978-3-030-59710-8

Scopus Subject Areas

Theoretical Computer Science
General Computer Science

User-Defined Keywords

Active learning
Deep reinforcement learning
Medical image classification

Access to Document

10.1007/978-3-030-59710-8_4

Cite this

Wang, J., Yan, Y., Zhang, Y., Cao, G., Yang, M., & Ng, M. K. (2020). Deep reinforcement active learning for medical image classification. In A. L. Martel, P. Abolmaesumi, D. Stoyanov, D. Mateus, M. A. Zuluaga, S. K. Zhou, D. Racoceanu, & L. Joskowicz (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2020: 23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part I (1st ed., pp. 33-42). (Lecture Notes in Computer Science; Vol. 12261), (Image Processing, Computer Vision, Pattern Recognition, and Graphics), (MICCAI: International Conference on Medical Image Computing and Computer-Assisted Intervention). Springer Cham. https://doi.org/10.1007/978-3-030-59710-8_4

Wang, Jingwen ; Yan, Yuguang ; Zhang, Yubing et al. / Deep reinforcement active learning for medical image classification. Medical Image Computing and Computer Assisted Intervention – MICCAI 2020: 23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part I. editor / Anne L. Martel ; Purang Abolmaesumi ; Danail Stoyanov ; Diana Mateus ; Maria A. Zuluaga ; S. Kevin Zhou ; Daniel Racoceanu ; Leo Joskowicz. 1st. ed. Springer Cham, 2020. pp. 33-42 (Lecture Notes in Computer Science). (Image Processing, Computer Vision, Pattern Recognition, and Graphics). (MICCAI: International Conference on Medical Image Computing and Computer-Assisted Intervention).

@inproceedings{2f0a97eac2f8458b84528f17f4d94203,

title = "Deep reinforcement active learning for medical image classification",

abstract = "In this paper, we propose a deep reinforcement learning algorithm for active learning on medical image data. Although deep learning has achieved great success on medical image processing, it relies on a large number of labeled data for training, which is expensive and time-consuming. Active learning, which follows a strategy to select and annotate informative samples, is an effective approach to alleviate this issue. However, most existing methods of active learning adopt a hand-design strategy, which cannot handle the dynamic procedure of classifier training. To address this issue, we model the procedure of active learning as a Markov decision process, and propose a deep reinforcement learning algorithm to learn a dynamic policy for active learning. To achieve this, we employ the actor-critic approach, and apply the deep deterministic policy gradient algorithm to train the model. We conduct experiments on two kinds of medical image data sets, and the results demonstrate that our method is able to learn better strategy compared with the existing hand-design ones.",

keywords = "Active learning, Deep reinforcement learning, Medical image classification",

author = "Jingwen Wang and Yuguang Yan and Yubing Zhang and Guiping Cao and Ming Yang and Ng, {Michael K.}",

note = "This work was supported by HKRGC GRF 12306616, 12200317, 12300218, 12300519, and 17201020. Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2020.; 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020 ; Conference date: 04-10-2020 Through 08-10-2020",

year = "2020",

month = sep,

day = "29",

doi = "10.1007/978-3-030-59710-8_4",

language = "English",

isbn = "9783030597092",

series = "Lecture Notes in Computer Science",

publisher = "Springer Cham",

pages = "33--42",

editor = "Martel, {Anne L.} and Purang Abolmaesumi and Danail Stoyanov and Diana Mateus and Zuluaga, {Maria A.} and Zhou, {S. Kevin} and Daniel Racoceanu and Leo Joskowicz",

booktitle = "Medical Image Computing and Computer Assisted Intervention – MICCAI 2020",

edition = "1st",

url = "https://link.springer.com/book/10.1007/978-3-030-59710-8",

}

Wang, J, Yan, Y, Zhang, Y, Cao, G, Yang, M & Ng, MK 2020, Deep reinforcement active learning for medical image classification. in AL Martel, P Abolmaesumi, D Stoyanov, D Mateus, MA Zuluaga, SK Zhou, D Racoceanu & L Joskowicz (eds), Medical Image Computing and Computer Assisted Intervention – MICCAI 2020: 23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part I. 1st edn, Lecture Notes in Computer Science, vol. 12261, Image Processing, Computer Vision, Pattern Recognition, and Graphics, MICCAI: International Conference on Medical Image Computing and Computer-Assisted Intervention, Springer Cham, pp. 33-42, 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020, Lima, Peru, 4/10/20. https://doi.org/10.1007/978-3-030-59710-8_4

Deep reinforcement active learning for medical image classification. / Wang, Jingwen; Yan, Yuguang; Zhang, Yubing et al.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2020: 23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part I. ed. / Anne L. Martel; Purang Abolmaesumi; Danail Stoyanov; Diana Mateus; Maria A. Zuluaga; S. Kevin Zhou; Daniel Racoceanu; Leo Joskowicz. 1st. ed. Springer Cham, 2020. p. 33-42 (Lecture Notes in Computer Science; Vol. 12261), (Image Processing, Computer Vision, Pattern Recognition, and Graphics), (MICCAI: International Conference on Medical Image Computing and Computer-Assisted Intervention).

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

TY - GEN

T1 - Deep reinforcement active learning for medical image classification

AU - Wang, Jingwen

AU - Yan, Yuguang

AU - Zhang, Yubing

AU - Cao, Guiping

AU - Yang, Ming

AU - Ng, Michael K.

N1 - This work was supported by HKRGC GRF 12306616, 12200317, 12300218, 12300519, and 17201020. Publisher Copyright: © Springer Nature Switzerland AG 2020.

PY - 2020/9/29

Y1 - 2020/9/29

N2 - In this paper, we propose a deep reinforcement learning algorithm for active learning on medical image data. Although deep learning has achieved great success on medical image processing, it relies on a large number of labeled data for training, which is expensive and time-consuming. Active learning, which follows a strategy to select and annotate informative samples, is an effective approach to alleviate this issue. However, most existing methods of active learning adopt a hand-design strategy, which cannot handle the dynamic procedure of classifier training. To address this issue, we model the procedure of active learning as a Markov decision process, and propose a deep reinforcement learning algorithm to learn a dynamic policy for active learning. To achieve this, we employ the actor-critic approach, and apply the deep deterministic policy gradient algorithm to train the model. We conduct experiments on two kinds of medical image data sets, and the results demonstrate that our method is able to learn better strategy compared with the existing hand-design ones.

AB - In this paper, we propose a deep reinforcement learning algorithm for active learning on medical image data. Although deep learning has achieved great success on medical image processing, it relies on a large number of labeled data for training, which is expensive and time-consuming. Active learning, which follows a strategy to select and annotate informative samples, is an effective approach to alleviate this issue. However, most existing methods of active learning adopt a hand-design strategy, which cannot handle the dynamic procedure of classifier training. To address this issue, we model the procedure of active learning as a Markov decision process, and propose a deep reinforcement learning algorithm to learn a dynamic policy for active learning. To achieve this, we employ the actor-critic approach, and apply the deep deterministic policy gradient algorithm to train the model. We conduct experiments on two kinds of medical image data sets, and the results demonstrate that our method is able to learn better strategy compared with the existing hand-design ones.

KW - Active learning

KW - Deep reinforcement learning

KW - Medical image classification

UR - http://www.scopus.com/inward/record.url?scp=85093085004&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-59710-8_4

DO - 10.1007/978-3-030-59710-8_4

M3 - Conference proceeding

AN - SCOPUS:85093085004

SN - 9783030597092

T3 - Lecture Notes in Computer Science

SP - 33

EP - 42

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2020

A2 - Martel, Anne L.

A2 - Abolmaesumi, Purang

A2 - Stoyanov, Danail

A2 - Mateus, Diana

A2 - Zuluaga, Maria A.

A2 - Zhou, S. Kevin

A2 - Racoceanu, Daniel

A2 - Joskowicz, Leo

PB - Springer Cham

T2 - 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020

Y2 - 4 October 2020 through 8 October 2020

ER -

Wang J, Yan Y, Zhang Y, Cao G, Yang M, Ng MK. Deep reinforcement active learning for medical image classification. In Martel AL, Abolmaesumi P, Stoyanov D, Mateus D, Zuluaga MA, Zhou SK, Racoceanu D, Joskowicz L, editors, Medical Image Computing and Computer Assisted Intervention – MICCAI 2020: 23rd International Conference, Lima, Peru, October 4–8, 2020, Proceedings, Part I. 1st ed. Springer Cham. 2020. p. 33-42. (Lecture Notes in Computer Science). (Image Processing, Computer Vision, Pattern Recognition, and Graphics). (MICCAI: International Conference on Medical Image Computing and Computer-Assisted Intervention). doi: 10.1007/978-3-030-59710-8_4

Deep reinforcement active learning for medical image classification

Abstract

Publication series

Conference

Scopus Subject Areas

User-Defined Keywords

Access to Document

Other files and links

Fingerprint

Cite this