Tooth numbering with polygonal segmentation on periapical radiographs: an artificial intelligence study.

Area of Science:

• * Dental Radiology
• * Artificial Intelligence in Medicine
• * Machine Learning Applications

Background:

• * Accurate tooth identification and numbering are critical for dental diagnostics and treatment planning.
• * Traditional methods of tooth identification can be time-consuming and prone to errors, especially for inexperienced clinicians.
• * Artificial intelligence (AI) offers potential solutions for automating and improving the accuracy of radiographic analysis.

Purpose of the Study:

• * To evaluate the efficacy of YOLOv5, an AI model, in automatically detecting and numbering teeth on periapical radiographs.
• * To validate the hypothesis that YOLOv5 can achieve high accuracy in this task.
• * To assess the potential of AI in enhancing dental radiographic interpretation.

Main Methods:

• * A dataset of 6,446 anonymized periapical radiographs was utilized.
• * Images were randomly allocated into training (80%), validation (10%), and testing (10%) sets.
• * Model performance was assessed using a confusion matrix and key performance metrics.

Main Results:

• * YOLOv5 achieved a true positive rate of 94.4% with a false positive rate of 4.4% and false negative rate of 1.2% during testing.
• * High recall (0.9882), precision (0.9548), and F1 scores (0.9712) indicate robust performance.
• * The model achieved an Area Under the Curve (AUC) of 0.603 on the ROC curve.

Conclusions:

• * YOLOv5 demonstrates near-perfect performance for tooth numbering in periapical radiography.
• * AI, like YOLOv5, can serve as a valuable guide for dentists, enhancing diagnostic accuracy and efficiency.
• * This technology has the potential to accelerate radiographic examination, reduce errors for less experienced dentists, and aid in dental education.