Deep learning tooth numbering on panoramic radiographs
A Convolutional Neural Network for Automatic Tooth Numbering in Panoramic Images
Authors: María Prados-Privado, Javier García Villalón, Antonio Blázquez Torres, Carlos Hugo Martínez-Martínez, Carlos Ivorra
Summary
This study presents a convolutional neural network (CNN) developed for automatic tooth numbering in panoramic radiographs using the FDI system.
A total of 8,000 anonymized images were collected from Asisa Dental S.A.U. centers in Madrid (Spain), curated by two experienced dentists. The model combines Matterport Mask R-CNN for object detection and ResNet-101 for classification, leveraging transfer learning from a previous model that achieved 99.24% accuracy in tooth detection.
The network was trained on 1,217 curated images (after filtering and quality control). Training involved 53 runs with 60–300 epochs each, varying learning rates between 0.0014 and 0.012.
The final model achieved:
Accuracy: 93.83% (total loss 6.17%)
Tooth Detection: 99.24% accuracy
Tooth Numbering: 93.83% accuracy
It correctly identified missing, filled, and metallic teeth in most clinical cases, though occasional numbering errors occurred for pontics and third molars.
The authors conclude that the model is reliable enough for use in clinical environments and demonstrates strong potential for automated diagnostic support.
Key Words
Tooth Numbering, Panoramic Radiography, Deep Learning, Mask RCNN, ResNet-101, Object Detection, Transfer Learning, Computer Vision, FDI System, Dental AI
Extracted Data
Year: 2021
Modality: Panoramic Radiography
Dataset: 8,000 images (total); 1,217 used for training and validation
Dataset Split: Not specified (53 training runs, cross-validated)
Network Architecture: Matterport Mask RCNN (backbone ResNet-101) + transfer learning from previous model
Metrics: Tooth Detection = 99.24%; Tooth Numbering = 93.83%; Total Loss = 6.17%
AP – Strategy: Manual annotation by two dentists using FDI system; cross-validation with Dice evaluation
AP – Professional Qty: 2 experienced general dentists
AP – Supervisor Presence: Yes (Asisa Dental Research Dept., Madrid)
AP – Experience Level: > 3 years clinical experience each
AP – Expertise Area: General Dentistry / Oral Radiology
AP – Tool or System: Mask RCNN + ResNet-101 (TensorFlow 1.14 / 2.2; AWS p3.8 instance with Tesla V100 GPUs)
Task: Object Detection + Classification (Tooth Numbering)
Project Objective: To automatically detect and number teeth in panoramic radiographs according to the FDI notation using deep learning methods
Clinical Relevance
Clinical importance: Automating tooth numbering can reduce diagnostic time and improve consistency in clinical documentation.
Innovation: By combining detection and numbering within a single CNN framework, this study shows the feasibility of AI-assisted dental charting directly from panoramic radiographs.
Practical application: The system achieves high reliability in real-world radiographs, including cases with metallic restorations and implants, making it a valuable support tool for dentists and radiologists.
