Objectives
This study aims to detect alveolar bone loss from dental panoramic radiography images using an artificial intelligence (AI) system.
Materials and Methods
- Dataset: 2276 panoramic radiography images (1137 with bone loss, 1139 periodontally healthy).
- Preprocessing: Images resized to 1472x718 pixels.
- AI System: Pretrained Google Net Inception v3 convolutional neural network (CNN) used for preprocessing and trained with transfer learning.
- Evaluation Metrics: Sensitivity, specificity, precision, accuracy, and F1 score calculated using a confusion matrix.
Results
- Detection Performance:
- Sensitivity: 0.94
- Specificity: 0.88
- Precision: 0.89
- Accuracy: 0.91
- F1 Score: 0.91
Conclusions
The CNN system successfully determines periodontal bone loss, indicating its potential use in facilitating diagnosis and treatment planning by oral physicians.
