Segmentation of lung nodule in CT images based on mask R-CNN

Menglu Liu; Junyu Dong; Xinghui Dong; Hui Yu; Lin Qi

doi:10.1109/ICAwST.2018.8517248

Segmentation of lung nodule in CT images based on mask R-CNN

Menglu Liu, Junyu Dong, Xinghui Dong, Hui Yu, Lin Qi

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

1478 Downloads (Pure)

Abstract

Due to the low-quality of CT images, the lack of annotated data, and the complex shapes of lung nodules, existing methods for lung nodules detection only predict the center of the nodule, whereas the nodule size is a very important diagnostic criteria but is neglected. In this paper, we employed the powerful object detection neural network “Mask R-CNN” for lung nodule segmentation, which provides contour information. Because of the imbalance between positive and negative samples, we trained classification networks based on block. We selected the classification network with the hightest accuracy. The selected classification network was used as the backbone of the image segmentation network—Mask R-CNN, which performs excellently on natural images. Lastly, Mask R-CNN model trained on the COCO data set was fine-tuned to segment pulmonary nodules. The model was tested on the LIDC-IDRI dataset.

Original language	English
Title of host publication	2018 9th International Conference on Awareness Science and Technology, iCAST 2018
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	95-100
Number of pages	6
ISBN (Electronic)	978-1-5386-5826-0
ISBN (Print)	978-1-5386-5827-7
DOIs	https://doi.org/10.1109/ICAwST.2018.8517248
Publication status	Published - 1 Nov 2018
Event	9th IEEE International Conference on Awareness Science and Technology - Fukuoka, Japan Duration: 19 Sept 2018 → 21 Sept 2018 http://www.design.kyushu-u.ac.jp/~icast/

Publication series

Name	IEEE iCAST Proceedings Series
Publisher	IEEE
ISSN (Print)	2325-5986
ISSN (Electronic)	2325-5994

Conference

Conference	9th IEEE International Conference on Awareness Science and Technology
Abbreviated title	iCAST 2018
Country/Territory	Japan
City	Fukuoka
Period	19/09/18 → 21/09/18
Internet address	http://www.design.kyushu-u.ac.jp/~icast/

Access to Document

10.1109/ICAwST.2018.8517248

Segmentation of Lung Nodule in CT Images_postprint
© © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Accepted author manuscript (Post-print), 0.98 MB

Cite this

@inproceedings{cfceafc33e7d455ab32ad8bbb5579596,

title = "Segmentation of lung nodule in CT images based on mask R-CNN",

abstract = "Due to the low-quality of CT images, the lack of annotated data, and the complex shapes of lung nodules, existing methods for lung nodules detection only predict the center of the nodule, whereas the nodule size is a very important diagnostic criteria but is neglected. In this paper, we employed the powerful object detection neural network “Mask R-CNN” for lung nodule segmentation, which provides contour information. Because of the imbalance between positive and negative samples, we trained classification networks based on block. We selected the classification network with the hightest accuracy. The selected classification network was used as the backbone of the image segmentation network—Mask R-CNN, which performs excellently on natural images. Lastly, Mask R-CNN model trained on the COCO data set was fine-tuned to segment pulmonary nodules. The model was tested on the LIDC-IDRI dataset.",

author = "Menglu Liu and Junyu Dong and Xinghui Dong and Hui Yu and Lin Qi",

year = "2018",

month = nov,

day = "1",

doi = "10.1109/ICAwST.2018.8517248",

language = "English",

isbn = "978-1-5386-5827-7",

series = "IEEE iCAST Proceedings Series",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "95--100",

booktitle = "2018 9th International Conference on Awareness Science and Technology, iCAST 2018",

address = "United States",

note = "9th IEEE International Conference on Awareness Science and Technology, iCAST 2018 ; Conference date: 19-09-2018 Through 21-09-2018",

url = "http://www.design.kyushu-u.ac.jp/\textasciitilde{}icast/",

}

Liu, M, Dong, J, Dong, X, Yu, H & Qi, L 2018, Segmentation of lung nodule in CT images based on mask R-CNN. in 2018 9th International Conference on Awareness Science and Technology, iCAST 2018., 8517248, IEEE iCAST Proceedings Series, Institute of Electrical and Electronics Engineers Inc., pp. 95-100, 9th IEEE International Conference on Awareness Science and Technology, Fukuoka, Japan, 19/09/18. https://doi.org/10.1109/ICAwST.2018.8517248

Segmentation of lung nodule in CT images based on mask R-CNN. / Liu, Menglu; Dong, Junyu; Dong, Xinghui et al.
2018 9th International Conference on Awareness Science and Technology, iCAST 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 95-100 8517248 (IEEE iCAST Proceedings Series).

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

TY - GEN

T1 - Segmentation of lung nodule in CT images based on mask R-CNN

AU - Liu, Menglu

AU - Dong, Junyu

AU - Dong, Xinghui

AU - Yu, Hui

AU - Qi, Lin

PY - 2018/11/1

Y1 - 2018/11/1

N2 - Due to the low-quality of CT images, the lack of annotated data, and the complex shapes of lung nodules, existing methods for lung nodules detection only predict the center of the nodule, whereas the nodule size is a very important diagnostic criteria but is neglected. In this paper, we employed the powerful object detection neural network “Mask R-CNN” for lung nodule segmentation, which provides contour information. Because of the imbalance between positive and negative samples, we trained classification networks based on block. We selected the classification network with the hightest accuracy. The selected classification network was used as the backbone of the image segmentation network—Mask R-CNN, which performs excellently on natural images. Lastly, Mask R-CNN model trained on the COCO data set was fine-tuned to segment pulmonary nodules. The model was tested on the LIDC-IDRI dataset.

AB - Due to the low-quality of CT images, the lack of annotated data, and the complex shapes of lung nodules, existing methods for lung nodules detection only predict the center of the nodule, whereas the nodule size is a very important diagnostic criteria but is neglected. In this paper, we employed the powerful object detection neural network “Mask R-CNN” for lung nodule segmentation, which provides contour information. Because of the imbalance between positive and negative samples, we trained classification networks based on block. We selected the classification network with the hightest accuracy. The selected classification network was used as the backbone of the image segmentation network—Mask R-CNN, which performs excellently on natural images. Lastly, Mask R-CNN model trained on the COCO data set was fine-tuned to segment pulmonary nodules. The model was tested on the LIDC-IDRI dataset.

U2 - 10.1109/ICAwST.2018.8517248

DO - 10.1109/ICAwST.2018.8517248

M3 - Conference contribution

SN - 978-1-5386-5827-7

T3 - IEEE iCAST Proceedings Series

SP - 95

EP - 100

BT - 2018 9th International Conference on Awareness Science and Technology, iCAST 2018

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 9th IEEE International Conference on Awareness Science and Technology

Y2 - 19 September 2018 through 21 September 2018

ER -

Segmentation of lung nodule in CT images based on mask R-CNN

Abstract

Publication series

Conference

Access to Document

Fingerprint

Cite this