Related Experiment Video
Updated: Jan 18, 2026
Author Spotlight: Advancing CBCT and Digital Dental Image Integration with AI-Assisted Digitization
Published on: February 23, 2024
Artificial intelligence model for application in dental traumatology
T Bani-Hani1, M Wedyan2, R Al-Fodeh3
1Division of Pediatric Dentistry, Preventive Dentistry Department, Faculty of Dentistry, Jordan University of Science and Technology, P.O.Box 3030, Irbid, 22110, Jordan. tgbanihani@just.edu.jo.
Background:
In recent years, healthcare systems have witnessed a tremendous advancement in diagnostic tools and technologies. The advent of artificial intelligence (AI) has enabled a paradigm shift in the practice of health sciences particularly in medicine. In the dental field, AI has been scarcely used in the various disciplines with no application in dental traumatology. This study proposes a deep-learning, convolutional neural networks (CNN)-based model for detection and classification of dental fractures.
Methods:
Plain periapical radiographs of injured teeth were retrieved from patients' records and annotated by two dentists trained in dental traumatology. The teeth were categorised into four groups: uncomplicated crown fractures, complicated crown fractures, crown-root fractures and root fractures. Data augmentation was done to enhance the power of the current dataset. Images were divided into training (80%) and test (20%) datasets. Python programming language was used to implement the CNN-based classification model. Cross validation was applied.
Results:
A total of 72 plain periapical radiographs of 108 fractured teeth were collected. The model achieved high accuracy in differentiating uncomplicated crown fractures from complicated ones (96.0%), from crown-root fractures (99.1%) and from root fractures (98.7%). Furthermore, the complicated injuries were distinguished from crown-root fractures and from root fractures with accuracy levels at 96.3% and 97.2% respectively. The model's overall accuracy in recognising the four classes was 78.7%.
Conclusion:
The proposed model showed excellent performance in the classification of dental fractures. The application of AI in paediatric dentistry, particularly in the field of dental trauma, is innovative and highly relevant to current trends in healthcare technology. Expansion of the current model to a larger dataset that includes the various types of injuries is recommended in future research. Such models can be a great asset for the less-experienced dentists in making accurate diagnosis and timely decisions. Future models employing panoramic radiographs could also help the medical practitioners at emergency services.
