A robust deep learning model for the classification of dental implant brands.
J Stomatol Oral Maxillofac Surg
; 125(5S1): 101818, 2024 Sep.
Article
en En
| MEDLINE
| ID: mdl-38462066
ABSTRACT
OBJECTIVE:
In cases where the brands of implants are not known, treatment options can be significantly limited in potential complications arising from implant procedures. This research aims to explore the application of deep learning techniques for the classification of dental implant systems using panoramic radiographs. The primary objective is to assess the superiority of the proposed model in achieving accurate and efficient dental implant classification. MATERIAL ANDMETHODS:
A comprehensive analysis was conducted using a diverse set of 25 convolutional neural network (CNN) models, including popular architectures such as VGG16, ResNet-50, EfficientNet, and ConvNeXt. The dataset of 1258 panoramic radiographs from patients who underwent implant treatment at faculty of dentistry was utilized for training and evaluation. Six different dental implant systems were employed as prototypes for the classification task. The precision, recall, F1 score, and support scores for each class have included in the classification accuracy report to ensure accurate and reliable results from the model.RESULTS:
The experimental results demonstrate that the proposed model consistently outperformed the other evaluated CNN architectures in terms of accuracy, precision, recall, and F1-score. With an impressive accuracy of 95.74 % and high precision and recall rates, the ConvNeXt model showcased its superiority in accurately classifying dental implant systems. Notably, the model's performance was achieved with a relatively smaller number of parameters, indicating its efficiency and speed during inference.CONCLUSION:
The findings highlight the effectiveness of deep learning techniques, particularly the proposed model, in accurately classifying dental implant systems from panoramic radiographs.Palabras clave
Texto completo:
1
Colección:
01-internacional
Banco de datos:
MEDLINE
Asunto principal:
Radiografía Panorámica
/
Implantes Dentales
/
Aprendizaje Profundo
Límite:
Humans
Idioma:
En
Revista:
J Stomatol Oral Maxillofac Surg
/
J. Stomatol. Oral Maxillofac. Surg
/
Journal of stomatology, oral and maxillofacial surgery (Online)
Año:
2024
Tipo del documento:
Article
País de afiliación:
Turquía