Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of Children
In this study, we aimed to develop and evaluate the performance of deep-learning models that automatically classify mesiodens in primary or mixed dentition panoramic radiographs. Panoramic radiographs of 550 patients with mesiodens and 550 patients without mesiodens were used. Primary or mixed denti...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-08-01
|
Series: | Diagnostics |
Subjects: | |
Online Access: | https://www.mdpi.com/2075-4418/11/8/1477 |
id |
doaj-7d7b7230e547442f83dcd9ae47572353 |
---|---|
record_format |
Article |
spelling |
doaj-7d7b7230e547442f83dcd9ae475723532021-08-26T13:40:33ZengMDPI AGDiagnostics2075-44182021-08-01111477147710.3390/diagnostics11081477Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of ChildrenYounghyun Ahn0Jae Joon Hwang1Yun-Hoa Jung2Taesung Jeong3Jonghyun Shin4Department of Pediatric Dentistry, School of Dentistry, Pusan National University, Yangsan 50612, KoreaDepartment of Oral and Maxillofacial Radiology, School of Dentistry, Pusan National University, Yangsan 50612, KoreaDepartment of Oral and Maxillofacial Radiology, School of Dentistry, Pusan National University, Yangsan 50612, KoreaDepartment of Pediatric Dentistry, School of Dentistry, Pusan National University, Yangsan 50612, KoreaDepartment of Pediatric Dentistry, School of Dentistry, Pusan National University, Yangsan 50612, KoreaIn this study, we aimed to develop and evaluate the performance of deep-learning models that automatically classify mesiodens in primary or mixed dentition panoramic radiographs. Panoramic radiographs of 550 patients with mesiodens and 550 patients without mesiodens were used. Primary or mixed dentition patients were included. SqueezeNet, ResNet-18, ResNet-101, and Inception-ResNet-V2 were each used to create deep-learning models. The accuracy, precision, recall, and F1 score of ResNet-101 and Inception-ResNet-V2 were higher than 90%. SqueezeNet exhibited relatively inferior results. In addition, we attempted to visualize the models using a class activation map. In images with mesiodens, the deep-learning models focused on the actual locations of the mesiodens in many cases. Deep-learning technologies may help clinicians with insufficient clinical experience in more accurate and faster diagnosis.https://www.mdpi.com/2075-4418/11/8/1477mesiodensartificial intelligencedeep learningconvolutional neural networks |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Younghyun Ahn Jae Joon Hwang Yun-Hoa Jung Taesung Jeong Jonghyun Shin |
spellingShingle |
Younghyun Ahn Jae Joon Hwang Yun-Hoa Jung Taesung Jeong Jonghyun Shin Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of Children Diagnostics mesiodens artificial intelligence deep learning convolutional neural networks |
author_facet |
Younghyun Ahn Jae Joon Hwang Yun-Hoa Jung Taesung Jeong Jonghyun Shin |
author_sort |
Younghyun Ahn |
title |
Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of Children |
title_short |
Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of Children |
title_full |
Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of Children |
title_fullStr |
Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of Children |
title_full_unstemmed |
Automated Mesiodens Classification System Using Deep Learning on Panoramic Radiographs of Children |
title_sort |
automated mesiodens classification system using deep learning on panoramic radiographs of children |
publisher |
MDPI AG |
series |
Diagnostics |
issn |
2075-4418 |
publishDate |
2021-08-01 |
description |
In this study, we aimed to develop and evaluate the performance of deep-learning models that automatically classify mesiodens in primary or mixed dentition panoramic radiographs. Panoramic radiographs of 550 patients with mesiodens and 550 patients without mesiodens were used. Primary or mixed dentition patients were included. SqueezeNet, ResNet-18, ResNet-101, and Inception-ResNet-V2 were each used to create deep-learning models. The accuracy, precision, recall, and F1 score of ResNet-101 and Inception-ResNet-V2 were higher than 90%. SqueezeNet exhibited relatively inferior results. In addition, we attempted to visualize the models using a class activation map. In images with mesiodens, the deep-learning models focused on the actual locations of the mesiodens in many cases. Deep-learning technologies may help clinicians with insufficient clinical experience in more accurate and faster diagnosis. |
topic |
mesiodens artificial intelligence deep learning convolutional neural networks |
url |
https://www.mdpi.com/2075-4418/11/8/1477 |
work_keys_str_mv |
AT younghyunahn automatedmesiodensclassificationsystemusingdeeplearningonpanoramicradiographsofchildren AT jaejoonhwang automatedmesiodensclassificationsystemusingdeeplearningonpanoramicradiographsofchildren AT yunhoajung automatedmesiodensclassificationsystemusingdeeplearningonpanoramicradiographsofchildren AT taesungjeong automatedmesiodensclassificationsystemusingdeeplearningonpanoramicradiographsofchildren AT jonghyunshin automatedmesiodensclassificationsystemusingdeeplearningonpanoramicradiographsofchildren |
_version_ |
1721194030882095104 |