An improved automatic system for aiding the detection of colon polyps using deep learning

Lishan Cai; Regina Beets-Tan; Sean Benson

doi:10.1109/BHI50953.2021.9508557

An improved automatic system for aiding the detection of colon polyps using deep learning

Lishan Cai, Regina Beets-Tan, Sean Benson

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceeding › Academic

Abstract

Colorectal cancer is responsible for the most cancer deaths after lung cancer. It has been well-established that early detection and removal of polyps can prevent colorectal cancer. It is therefore essential that automated polyp detection has the highest sensitivity and precision possible in order to detect the most cases and prevent unnecessary treatment. We present a deep learning model based on YOLOv3 that was trained to detect polyps. Training made use of the 39308 images of 78 polyps and 393 completely healthy images from the SUN database. The model was subsequently validated using both the public CVC-clinic and ETIS-Larib datasets containing both standard defintion (SD) and high definition (HD) images. The per-image polyp detection sensitivity(precision) was calculated as 91.5(96.6)% and 86.5(94.2)% for the CVC-clinic and Etis-Larib datasets, respectively. These results represent the best-known performance in the validation datasets in comparison with the results of a recent review.

Original language	English
Title of host publication	BHI 2021 - 2021 IEEE EMBS International Conference on Biomedical and Health Informatics, Proceedings
Publisher	IEEE
ISBN (Electronic)	9781665403580
DOIs	https://doi.org/10.1109/BHI50953.2021.9508557
Publication status	Published - 1 Jan 2021
Event	IEEE EMBS International Conference on Biomedical and Health Informatics 2021 - Online, Greece Duration: 27 Jul 2021 → 30 Jul 2021

Conference

Conference	IEEE EMBS International Conference on Biomedical and Health Informatics 2021
Abbreviated title	BHI 2021
Country/Territory	Greece
Period	27/07/21 → 30/07/21

Keywords

Artificial intelligence
Colonoscopy
Deep learning
Polyp detection
YOLOv3

Access to Document

10.1109/BHI50953.2021.9508557Licence: CC BY

Cite this

@inproceedings{de06b1ed7af6484dbe54ad6debc3e7a0,

title = "An improved automatic system for aiding the detection of colon polyps using deep learning",

abstract = "Colorectal cancer is responsible for the most cancer deaths after lung cancer. It has been well-established that early detection and removal of polyps can prevent colorectal cancer. It is therefore essential that automated polyp detection has the highest sensitivity and precision possible in order to detect the most cases and prevent unnecessary treatment. We present a deep learning model based on YOLOv3 that was trained to detect polyps. Training made use of the 39308 images of 78 polyps and 393 completely healthy images from the SUN database. The model was subsequently validated using both the public CVC-clinic and ETIS-Larib datasets containing both standard defintion (SD) and high definition (HD) images. The per-image polyp detection sensitivity(precision) was calculated as 91.5(96.6)% and 86.5(94.2)% for the CVC-clinic and Etis-Larib datasets, respectively. These results represent the best-known performance in the validation datasets in comparison with the results of a recent review.",

keywords = "Artificial intelligence, Colonoscopy, Deep learning, Polyp detection, YOLOv3",

author = "Lishan Cai and Regina Beets-Tan and Sean Benson",

note = "Funding Information: This project has received funding from the European Union{\textquoteright}s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 857894. Funding Information: This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sk?odowska-Curie grant agreement No 857894. Publisher Copyright: {\textcopyright} 2021 IEEE; IEEE EMBS International Conference on Biomedical and Health Informatics 2021, BHI 2021 ; Conference date: 27-07-2021 Through 30-07-2021",

year = "2021",

month = jan,

day = "1",

doi = "10.1109/BHI50953.2021.9508557",

language = "English",

booktitle = "BHI 2021 - 2021 IEEE EMBS International Conference on Biomedical and Health Informatics, Proceedings",

publisher = "IEEE",

address = "United States",

}

Cai, L , Beets-Tan, R & Benson, S 2021, An improved automatic system for aiding the detection of colon polyps using deep learning. in BHI 2021 - 2021 IEEE EMBS International Conference on Biomedical and Health Informatics, Proceedings. IEEE, IEEE EMBS International Conference on Biomedical and Health Informatics 2021, Greece, 27/07/21. https://doi.org/10.1109/BHI50953.2021.9508557

TY - GEN

T1 - An improved automatic system for aiding the detection of colon polyps using deep learning

AU - Cai, Lishan

AU - Beets-Tan, Regina

AU - Benson, Sean

N1 - Funding Information: This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 857894. Funding Information: This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sk?odowska-Curie grant agreement No 857894. Publisher Copyright: © 2021 IEEE

PY - 2021/1/1

Y1 - 2021/1/1

N2 - Colorectal cancer is responsible for the most cancer deaths after lung cancer. It has been well-established that early detection and removal of polyps can prevent colorectal cancer. It is therefore essential that automated polyp detection has the highest sensitivity and precision possible in order to detect the most cases and prevent unnecessary treatment. We present a deep learning model based on YOLOv3 that was trained to detect polyps. Training made use of the 39308 images of 78 polyps and 393 completely healthy images from the SUN database. The model was subsequently validated using both the public CVC-clinic and ETIS-Larib datasets containing both standard defintion (SD) and high definition (HD) images. The per-image polyp detection sensitivity(precision) was calculated as 91.5(96.6)% and 86.5(94.2)% for the CVC-clinic and Etis-Larib datasets, respectively. These results represent the best-known performance in the validation datasets in comparison with the results of a recent review.

AB - Colorectal cancer is responsible for the most cancer deaths after lung cancer. It has been well-established that early detection and removal of polyps can prevent colorectal cancer. It is therefore essential that automated polyp detection has the highest sensitivity and precision possible in order to detect the most cases and prevent unnecessary treatment. We present a deep learning model based on YOLOv3 that was trained to detect polyps. Training made use of the 39308 images of 78 polyps and 393 completely healthy images from the SUN database. The model was subsequently validated using both the public CVC-clinic and ETIS-Larib datasets containing both standard defintion (SD) and high definition (HD) images. The per-image polyp detection sensitivity(precision) was calculated as 91.5(96.6)% and 86.5(94.2)% for the CVC-clinic and Etis-Larib datasets, respectively. These results represent the best-known performance in the validation datasets in comparison with the results of a recent review.

KW - Artificial intelligence

KW - Colonoscopy

KW - Deep learning

KW - Polyp detection

KW - YOLOv3

U2 - 10.1109/BHI50953.2021.9508557

DO - 10.1109/BHI50953.2021.9508557

M3 - Conference article in proceeding

BT - BHI 2021 - 2021 IEEE EMBS International Conference on Biomedical and Health Informatics, Proceedings

PB - IEEE

T2 - IEEE EMBS International Conference on Biomedical and Health Informatics 2021

Y2 - 27 July 2021 through 30 July 2021

ER -

An improved automatic system for aiding the detection of colon polyps using deep learning

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this