Skip to content

Deep neural networks show a superior performance to all 42 dermatologists in onychomycosis diagnosis

 

We created datasets of standardized nail images using a region-based convolutional neural network (R-CNN) trained to distinguish a nail from a background. We used R-CNN to generate training datasets of 49,567 images, which were then used to fine-tune the ResNet-152 and VGG-19 models.

The validation datasets comprised 100 and 194 images from Inje University (B1 and B2 datasets, respectively), 125 images from Hallym University (C dataset), and 939 images from Seoul National University (D dataset).


The results from the AI ensemble model (ResNet-152 + VGG-19 + feedforward neural networks) showed the test sensitivity/specificity/ area under the curve values of (96.0 / 94.7 / 0.98), (82.7 / 96.7 / 0.95), (92.3 / 79.3 / 0.93), (87.7 / 69.3 / 0.82) for the B1, B2, C, and D datasets, respectively.

With the combination of the B1 and C datasets, the AI Youden index was significantly higher than that of 42 dermatologists who performed the same assessment manually (p = 0.01). For the combinations of B1+C and B2+D dataset, almost none of the dermatologists performed as well as the AI.


The deep learning model trained with the dataset comprising 49,567 images, achieved a diagnostic accuracy for the classification of onychomycosis, which was superior to that of most of the dermatologists who participated in this study.

 

번호 제목 글쓴이 날짜 조회 수
1671 git 특정 버젼 WHRIA 2020.07.11 34
1670 naver rexnet clova AI [1] WHRIA 2020.07.04 228
1669 sql python WHRIA 2020.07.01 180
1668 modprobe WHRIA 2020.06.29 6432
1667 tab 구조 html WHRIA 2020.06.28 79
1666 imagenet classification SOTA WHRIA 2020.06.27 1398
1665 할일 WHRIA 2020.06.24 1464
1664 caffe hdf5 error WHRIA 2020.06.22 187
1663 resnest50 caffe WHRIA 2020.06.21 163
1662 reddit downloader WHRIA 2020.06.21 204
1661 바이두 WHRIA 2020.06.20 627
1660 regression model [6] WHRIA 2020.06.18 167
1659 MS bot framework [3] WHRIA 2020.06.13 99
1658 BERT WHRIA 2020.06.10 3223
1657 영어로 숫자 표현 WHRIA 2020.05.28 4447

Powered by Xpress Engine / Designed by Sketchbook

sketchbook5, 스케치북5

sketchbook5, 스케치북5

나눔글꼴 설치 안내


이 PC에는 나눔글꼴이 설치되어 있지 않습니다.

이 사이트를 나눔글꼴로 보기 위해서는
나눔글꼴을 설치해야 합니다.

설치 취소