Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical Radiographs
In the absence of accurate medical records, it is critical to correctly classify implant fixture systems using periapical radiographs to provide accurate diagnoses and treatments to patients or to respond to complications. The purpose of this study was to evaluate whether deep neural networks can id...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-04-01
|
| Series: | Journal of Clinical Medicine |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2077-0383/9/4/1117 |
| _version_ | 1797570655629082624 |
|---|---|
| author | Jong-Eun Kim Na-Eun Nam June-Sung Shim Yun-Hoa Jung Bong-Hae Cho Jae Joon Hwang |
| author_facet | Jong-Eun Kim Na-Eun Nam June-Sung Shim Yun-Hoa Jung Bong-Hae Cho Jae Joon Hwang |
| author_sort | Jong-Eun Kim |
| collection | DOAJ |
| description | In the absence of accurate medical records, it is critical to correctly classify implant fixture systems using periapical radiographs to provide accurate diagnoses and treatments to patients or to respond to complications. The purpose of this study was to evaluate whether deep neural networks can identify four different types of implants on intraoral radiographs. In this study, images of 801 patients who underwent periapical radiographs between 2005 and 2019 at Yonsei University Dental Hospital were used. Images containing the following four types of implants were selected: Brånemark Mk TiUnite, Dentium Implantium, Straumann Bone Level, and Straumann Tissue Level. SqueezeNet, GoogLeNet, ResNet-18, MobileNet-v2, and ResNet-50 were tested to determine the optimal pre-trained network architecture. The accuracy, precision, recall, and F1 score were calculated for each network using a confusion matrix. All five models showed a test accuracy exceeding 90%. SqueezeNet and MobileNet-v2, which are small networks with less than four million parameters, showed an accuracy of approximately 96% and 97%, respectively. The results of this study confirmed that convolutional neural networks can classify the four implant fixtures with high accuracy even with a relatively small network and a small number of images. This may solve the inconveniences associated with unnecessary treatments and medical expenses caused by lack of knowledge about the exact type of implant. |
| first_indexed | 2024-03-10T20:28:09Z |
| format | Article |
| id | doaj.art-7310646b9c69425ea1064de40afe19c9 |
| institution | Directory Open Access Journal |
| issn | 2077-0383 |
| language | English |
| last_indexed | 2024-03-10T20:28:09Z |
| publishDate | 2020-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Journal of Clinical Medicine |
| spelling | doaj.art-7310646b9c69425ea1064de40afe19c92023-11-19T21:37:04ZengMDPI AGJournal of Clinical Medicine2077-03832020-04-0194111710.3390/jcm9041117Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical RadiographsJong-Eun Kim0Na-Eun Nam1June-Sung Shim2Yun-Hoa Jung3Bong-Hae Cho4Jae Joon Hwang5Department of Prosthodontics, Yonsei University College of Dentistry, Yonsei-ro 50-1, Seodaemun-gu, Seoul 03722, KoreaDepartment of Prosthodontics, Yonsei University College of Dentistry, Yonsei-ro 50-1, Seodaemun-gu, Seoul 03722, KoreaDepartment of Prosthodontics, Yonsei University College of Dentistry, Yonsei-ro 50-1, Seodaemun-gu, Seoul 03722, KoreaDepartment of Oral and Maxillofacial Radiology, School of Dentistry, Pusan National University, Dental Research Institute, Yangsan 50610, KoreaDepartment of Oral and Maxillofacial Radiology, School of Dentistry, Pusan National University, Dental Research Institute, Yangsan 50610, KoreaDepartment of Oral and Maxillofacial Radiology, School of Dentistry, Pusan National University, Dental Research Institute, Yangsan 50610, KoreaIn the absence of accurate medical records, it is critical to correctly classify implant fixture systems using periapical radiographs to provide accurate diagnoses and treatments to patients or to respond to complications. The purpose of this study was to evaluate whether deep neural networks can identify four different types of implants on intraoral radiographs. In this study, images of 801 patients who underwent periapical radiographs between 2005 and 2019 at Yonsei University Dental Hospital were used. Images containing the following four types of implants were selected: Brånemark Mk TiUnite, Dentium Implantium, Straumann Bone Level, and Straumann Tissue Level. SqueezeNet, GoogLeNet, ResNet-18, MobileNet-v2, and ResNet-50 were tested to determine the optimal pre-trained network architecture. The accuracy, precision, recall, and F1 score were calculated for each network using a confusion matrix. All five models showed a test accuracy exceeding 90%. SqueezeNet and MobileNet-v2, which are small networks with less than four million parameters, showed an accuracy of approximately 96% and 97%, respectively. The results of this study confirmed that convolutional neural networks can classify the four implant fixtures with high accuracy even with a relatively small network and a small number of images. This may solve the inconveniences associated with unnecessary treatments and medical expenses caused by lack of knowledge about the exact type of implant.https://www.mdpi.com/2077-0383/9/4/1117implant fixture classificationartificial intelligencedeep learningconvolutional neural networksperiapical radiographs |
| spellingShingle | Jong-Eun Kim Na-Eun Nam June-Sung Shim Yun-Hoa Jung Bong-Hae Cho Jae Joon Hwang Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical Radiographs Journal of Clinical Medicine implant fixture classification artificial intelligence deep learning convolutional neural networks periapical radiographs |
| title | Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical Radiographs |
| title_full | Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical Radiographs |
| title_fullStr | Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical Radiographs |
| title_full_unstemmed | Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical Radiographs |
| title_short | Transfer Learning via Deep Neural Networks for Implant Fixture System Classification Using Periapical Radiographs |
| title_sort | transfer learning via deep neural networks for implant fixture system classification using periapical radiographs |
| topic | implant fixture classification artificial intelligence deep learning convolutional neural networks periapical radiographs |
| url | https://www.mdpi.com/2077-0383/9/4/1117 |
| work_keys_str_mv | AT jongeunkim transferlearningviadeepneuralnetworksforimplantfixturesystemclassificationusingperiapicalradiographs AT naeunnam transferlearningviadeepneuralnetworksforimplantfixturesystemclassificationusingperiapicalradiographs AT junesungshim transferlearningviadeepneuralnetworksforimplantfixturesystemclassificationusingperiapicalradiographs AT yunhoajung transferlearningviadeepneuralnetworksforimplantfixturesystemclassificationusingperiapicalradiographs AT bonghaecho transferlearningviadeepneuralnetworksforimplantfixturesystemclassificationusingperiapicalradiographs AT jaejoonhwang transferlearningviadeepneuralnetworksforimplantfixturesystemclassificationusingperiapicalradiographs |