Automated Detection of Osteoporosis in Dental Panoramic Images Using a ResNet18-Based Convolutional Neural Network

N. Sánchez Martín ^[1] ; V. Vera González ^[2] ; F. Panetsos Petrova ^[2] ; G. Kontaxakis ^[1]
1. [1] Universidad Politécnica de Madrid
  
  Universidad Politécnica de Madrid
  
  Madrid, España
2. [2] Universidad Complutense de Madrid
  
  Universidad Complutense de Madrid
  
  Madrid, España
Localización: CASEIB 2024. Libro de Actas del XLII Congreso Anual de la Sociedad Española de Ingeniería Biomédica, 2024, ISBN 978-84-09-67332-2, págs. 345-348
Idioma: inglés
Enlaces
- Texto Completo Libro
Resumen
- Osteoporosis significantly affects the elderly population, leading to an increased risk of fractures and higher healthcare costs, making early detection crucial for improving patient outcomes. In this study, a robust convolutional neural network was developed for the automated detection of osteoporosis using dental panoramic images. Image quality was enhanced through resizing, manual selection of regions of interest, and fuzzy c-means clustering. To improve model generalization, data augmentation techniques were applied, and the ResNet18 architecture, pre-trained on ImageNet, was employed for binary classification. The model achieved an overall accuracy of 84.85% and a precision of 93.33% in predicting osteoporosis, demonstrating the potential of deep learning for the automated detection of osteoporosis using dental images, which are both cost-effective and routinely captured during dental examinations.