Dental Age-Group Classification from Panoramic Radiographs Using Convolutional Neural Networks

Area of Science:

• Radiology and Forensic Odontology
• Artificial Intelligence in Healthcare

Background:

• Accurate age determination is crucial for legal, forensic, and clinical applications.
• Conventional dental age estimation methods lack reliability in adults and seniors.
• Dental structures provide reliable age-related changes throughout life.

Purpose of the Study:

• To evaluate an automated deep learning approach for age-group classification in adults using panoramic dental radiographs.
• To compare a custom Convolutional Neural Network (CNN) with pre-trained deep learning models for this task.

Main Methods:

• Analysis of 1469 panoramic dental radiographs from Egyptian individuals (25-70 years).
• Classification into five age categories using a custom CNN and established deep learning architectures.
• Training models to identify age-related patterns in dental radiographs.

Main Results:

• The custom CNN achieved the highest accuracy (85.2%), outperforming other models like YOLOv8 (79.1%).
• The custom CNN demonstrated the lowest prediction error (MAE = 1.92 years; RMSE = 5.46 years).
• Deep learning models showed strong performance in adult and senior age-group classification.

Conclusions:

• Deep learning analysis of dental radiographs shows promise as a supportive tool for adult age-group classification.
• This AI-driven method can complement traditional age assessment techniques.
• Further validation across diverse populations is needed for broader applicability.