한빛사 논문
Yunseob Hwang,1,2 Han Hee Lee,4,5 Chunghyun Park,1 Bayu Adhi Tama,1 Jin Su Kim,4 Dae Young Cheung,4 Woo Chul Chung,4 Young-Seok Cho,4 Kang-Moon Lee,4 Myung-Gyu Choi,4 Seungchul Lee1,2,3,* and Bo-In Lee4,*
1Department of Mechanical Engineering, 2Postech-Catholic Biomedical Engineering Institute, 3Graduate School of Artificial Intelligence, Pohang University of Science and Technology (POSTECH), Pohang, 4Division of Gastroenterology, Department of Internal Medicine and 5Postech-Catholic Biomedical Engineering Institute, College of Medicine, The Catholic University of Korea, Seoul, Korea
*Corresponding:
Seungchul Lee, Department of Mechanical Engineering and Graduate School of Artificial Intelligence, Pohang University of Science and Technology (POSTECH), 5th Engineering Building, 77 Chengam-Ro, Nam-Gu Pohang, Gyeongbuk 37673, Korea.
Bo-InLee, Division of Gastroenterology, Department of Internal Medicine, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, 222 Banpo-daero, Seocho-gu, Seoul 137-701, Korea.
The authors Yunseob Hwang and Han Hee Lee contributedequally to this work as the first authors.
Abstract
Background
Although great advances in artificial intelligence for interpreting small bowel capsule endoscopy (SBCE) images have been made in recent years, its practical use is still limited. The aim of this study was to develop a more practical convolutional neural network (CNN) algorithm for the automatic detection of various small bowel lesions.
Methods
A total of 7556 images were collected for the training dataset from 526 SBCE videos. Abnormal images were classified into two categories: hemorrhagic lesions (red spot/angioectasia/active bleeding) and ulcerative lesions (erosion/ulcer/stricture). A CNN algorithm based on VGGNet was trained in two different ways: the combined model (hemorrhagic and ulcerative lesions trained separately) and the binary model (all abnormal images trained without discrimination). The detected lesions were visualized using a gradient class activation map (Grad‐CAM). The two models were validated using 5,760 independent images taken at two other academic hospitals.
Results
Both the combined and binary models acquired high accuracy for lesion detection, and the difference between the two models was not significant (96.83% vs 96.62%, P = 0.122). However, the combined model showed higher sensitivity (97.61% vs 95.07%, P < 0.001) and higher accuracy for individual lesions from the hemorrhagic and ulcerative categories than the binary model. The combined model also revealed more accurate localization of the culprit area on images evaluated by the Grad‐CAM.
Conclusions
Diagnostic sensitivity and classification of small bowel lesions using a convolutional neural network are improved by the independent training for hemorrhagic and ulcerative lesions. Grad‐CAM is highly effective in localizing the lesions.
논문정보
관련 링크
관련분야 연구자보기
소속기관 논문보기
관련분야 논문보기