AI-based prediction and validation of vertical angulation errors in intraoral periapical radiography

Objectives: This study aimed to develop and validate an artificial intelligence (AI)-based software for accurate prediction and correction of vertical angulation errors in intraoral periapical radiography, thereby improving diagnostic accuracy and imaging efficiency. Methods: This study employed a two-phase approach. In the development phase, a dataset of intraoral periapical radiographs with a radiopaque marker placed at varying vertical angulations was collected. Angulation errors were calculated and used to train a custom AI model for error prediction. In the validation phase, the trained AI software analyzed a new set of radiographs, and its predictions were compared with a gold standard reference radiograph obtained using a position indicating device (PID). Statistical analysis was performed to assess the correlation and agreement between AI predictions and PID-based measurements. Results: The AI model demonstrated high accuracy in predicting vertical angulation errors. A strong correlation was observed between AI-predicted values and PID-based measurements, indicating reliable performance. The model showed potential to achieve precision comparable to, or exceeding, conventional methods, while offering advantages in speed and automation. Conclusions: The developed AI-based system represents a promising advancement in dental radiography by enabling real-time detection and correction of angulation errors. Its implementation in clinical practice could reduce the need for repeat exposures, minimize patient radiation dose, and enhance workflow efficiency. This technology has the potential to establish an automated standard for optimizing radiographic quality and improving diagnostic outcomes.