A Convolutional Neural Network for Automatic Tooth Numbering in Panoramic Images

Area of Science:

• Dentistry
• Artificial Intelligence
• Medical Imaging

Background:

• Accurate tooth numbering is crucial for dental diagnosis and treatment planning.
• Manual interpretation of tooth numbering in panoramic radiographs can be time-consuming and prone to errors.
• Automating this process can enhance diagnostic efficiency and consistency.

Purpose of the Study:

• To develop and evaluate a convolutional neural network (CNN) for automatic tooth numbering in panoramic dental radiographs.
• To improve the accuracy and speed of tooth identification and numbering in clinical dental practice.

Main Methods:

• A dataset of 8,000 panoramic radiographs was annotated by experienced dentists.
• A two-layer neural network architecture was employed, integrating object detection (Matterport Mask RCNN) and classification (ResNet101).
• Transfer learning techniques were utilized to optimize computing time and precision.

Main Results:

• The proposed CNN model achieved an overall accuracy of 93.83% with a total loss of 6.17%.
• The architecture demonstrated high performance with 99.24% accuracy in tooth detection.
• The model accurately numbered teeth in various oral health conditions with 93.83% accuracy.

Conclusions:

• The developed CNN offers a reliable and accurate automated solution for tooth numbering in panoramic radiographs.
• This AI-driven approach has the potential to significantly streamline the diagnostic workflow in dentistry.
• The model's high accuracy in detection and numbering supports its clinical applicability for improving dental diagnostics.