Oral mucosal lesions triage via YOLOv7 models

Background/purpose: The global incidence of lip and oral cavity cancer continues to rise, necessitating improved early detection methods. This study leverages the capabilities of computer vision and deep learning to enhance the early detection and classification of oral mucosal lesions. Methods: A d...

Full description

Saved in:
Bibliographic Details
Main Authors: Yu Hsu, Cheng-Ying Chou, Yu-Cheng Huang, Yu-Chieh Liu, Yong-Long Lin, Zi-Ping Zhong, Jun-Kai Liao, Jun-Ching Lee, Hsin-Yu Chen, Jang-Jaer Lee, Shyh-Jye Chen
Format: Article
Language:English
Published: Elsevier 2025-07-01
Series:Journal of the Formosan Medical Association
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S0929664624003139
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850169952917520384
author Yu Hsu
Cheng-Ying Chou
Yu-Cheng Huang
Yu-Chieh Liu
Yong-Long Lin
Zi-Ping Zhong
Jun-Kai Liao
Jun-Ching Lee
Hsin-Yu Chen
Jang-Jaer Lee
Shyh-Jye Chen
author_facet Yu Hsu
Cheng-Ying Chou
Yu-Cheng Huang
Yu-Chieh Liu
Yong-Long Lin
Zi-Ping Zhong
Jun-Kai Liao
Jun-Ching Lee
Hsin-Yu Chen
Jang-Jaer Lee
Shyh-Jye Chen
author_sort Yu Hsu
collection DOAJ
description Background/purpose: The global incidence of lip and oral cavity cancer continues to rise, necessitating improved early detection methods. This study leverages the capabilities of computer vision and deep learning to enhance the early detection and classification of oral mucosal lesions. Methods: A dataset initially consisting of 6903 white-light macroscopic images collected from 2006 to 2013 was expanded to over 50,000 images to train the YOLOv7 deep learning model. Lesions were categorized into three referral grades: benign (green), potentially malignant (yellow), and malignant (red), facilitating efficient triage. Results: The YOLOv7 models, particularly the YOLOv7-E6, demonstrated high precision and recall across all lesion categories. The YOLOv7-D6 model excelled at identifying malignant lesions with notable precision, recall, and F1 scores. Enhancements, including the integration of coordinate attention in the YOLOv7-D6-CA model, significantly improved the accuracy of lesion classification. Conclusion: The study underscores the robust comparison of various YOLOv7 model configurations in the classification to triage oral lesions. The overall results highlight the potential of deep learning models to contribute to the early detection of oral cancers, offering valuable tools for both clinical settings and remote screening applications.
format Article
id doaj-art-27cf55c27e2c467a8dc71d0079c97b3f
institution OA Journals
issn 0929-6646
language English
publishDate 2025-07-01
publisher Elsevier
record_format Article
series Journal of the Formosan Medical Association
spelling doaj-art-27cf55c27e2c467a8dc71d0079c97b3f2025-08-20T02:20:37ZengElsevierJournal of the Formosan Medical Association0929-66462025-07-01124762162710.1016/j.jfma.2024.07.010Oral mucosal lesions triage via YOLOv7 modelsYu Hsu0Cheng-Ying Chou1Yu-Cheng Huang2Yu-Chieh Liu3Yong-Long Lin4Zi-Ping Zhong5Jun-Kai Liao6Jun-Ching Lee7Hsin-Yu Chen8Jang-Jaer Lee9Shyh-Jye Chen10Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan; Graduate Institute of Clinical Medicine, College of Medicine, National Taiwan University, Taipei, TaiwanDepartment of Biomechatronics Engineering, National Taiwan University, Taipei, TaiwanDepartment of Medical Imaging, National Taiwan University Hospital, Taipei, TaiwanDepartment of Biomechatronics Engineering, National Taiwan University, Taipei, TaiwanDepartment of Biomechatronics Engineering, National Taiwan University, Taipei, TaiwanDepartment of Biomechatronics Engineering, National Taiwan University, Taipei, TaiwanDepartment of Biomechatronics Engineering, National Taiwan University, Taipei, TaiwanDepartment of Dentistry, National Taiwan University Hospital, Taipei, TaiwanDepartment of Dentistry, National Taiwan University Hospital, Taipei, TaiwanDepartment of Dentistry, National Taiwan University Hospital, Taipei, Taiwan; Department of Dentistry, College of Medicine, National Taiwan University, Taipei, Taiwan; Corresponding author. Department of Dentistry, National Taiwan University Hospital, No. 7, Chung-Shan S. Rd., Taipei City, 10002, Taiwan.Department of Medical Imaging, National Taiwan University Hospital, Taipei, Taiwan; Graduate Institute of Clinical Medicine, College of Medicine, National Taiwan University, Taipei, Taiwan; Department of Radiology, College of Medicine, National Taiwan University, Taipei, Taiwan; Corresponding author. Department of Medical Imaging, National Taiwan University Hospital, No. 7, Chung-Shan S. Rd., Taipei City, 10002, Taiwan.Background/purpose: The global incidence of lip and oral cavity cancer continues to rise, necessitating improved early detection methods. This study leverages the capabilities of computer vision and deep learning to enhance the early detection and classification of oral mucosal lesions. Methods: A dataset initially consisting of 6903 white-light macroscopic images collected from 2006 to 2013 was expanded to over 50,000 images to train the YOLOv7 deep learning model. Lesions were categorized into three referral grades: benign (green), potentially malignant (yellow), and malignant (red), facilitating efficient triage. Results: The YOLOv7 models, particularly the YOLOv7-E6, demonstrated high precision and recall across all lesion categories. The YOLOv7-D6 model excelled at identifying malignant lesions with notable precision, recall, and F1 scores. Enhancements, including the integration of coordinate attention in the YOLOv7-D6-CA model, significantly improved the accuracy of lesion classification. Conclusion: The study underscores the robust comparison of various YOLOv7 model configurations in the classification to triage oral lesions. The overall results highlight the potential of deep learning models to contribute to the early detection of oral cancers, offering valuable tools for both clinical settings and remote screening applications.http://www.sciencedirect.com/science/article/pii/S0929664624003139Mouth neoplasmsOral pathologyDeep learningArtificial intelligenceComputer assisted decision making
spellingShingle Yu Hsu
Cheng-Ying Chou
Yu-Cheng Huang
Yu-Chieh Liu
Yong-Long Lin
Zi-Ping Zhong
Jun-Kai Liao
Jun-Ching Lee
Hsin-Yu Chen
Jang-Jaer Lee
Shyh-Jye Chen
Oral mucosal lesions triage via YOLOv7 models
Journal of the Formosan Medical Association
Mouth neoplasms
Oral pathology
Deep learning
Artificial intelligence
Computer assisted decision making
title Oral mucosal lesions triage via YOLOv7 models
title_full Oral mucosal lesions triage via YOLOv7 models
title_fullStr Oral mucosal lesions triage via YOLOv7 models
title_full_unstemmed Oral mucosal lesions triage via YOLOv7 models
title_short Oral mucosal lesions triage via YOLOv7 models
title_sort oral mucosal lesions triage via yolov7 models
topic Mouth neoplasms
Oral pathology
Deep learning
Artificial intelligence
Computer assisted decision making
url http://www.sciencedirect.com/science/article/pii/S0929664624003139
work_keys_str_mv AT yuhsu oralmucosallesionstriageviayolov7models
AT chengyingchou oralmucosallesionstriageviayolov7models
AT yuchenghuang oralmucosallesionstriageviayolov7models
AT yuchiehliu oralmucosallesionstriageviayolov7models
AT yonglonglin oralmucosallesionstriageviayolov7models
AT zipingzhong oralmucosallesionstriageviayolov7models
AT junkailiao oralmucosallesionstriageviayolov7models
AT junchinglee oralmucosallesionstriageviayolov7models
AT hsinyuchen oralmucosallesionstriageviayolov7models
AT jangjaerlee oralmucosallesionstriageviayolov7models
AT shyhjyechen oralmucosallesionstriageviayolov7models