AUTOMATED DETECTION AND CLASSIFICATION OF DENTAL ANOMALIES FROM OPG X-RAYS USING YOLOV10

Abstract

The accurate detection and classification of dental anomalies from radiographic images play a crucial role in modern dentistry for effective diagnosis and treatment planning. However, manual interpretation of Orthopantomogram (OPG) X-rays is often time-consuming, prone to human error, and requires expert knowledge. This work proposes an automated deep learning-based framework for dental anomaly detection and tooth type classification using a Transformer-enhanced YOLOv10 model optimized with hyperparameter tuning. The system is trained on a labeled dental OPG dataset containing six classes, including four tooth types (molar, premolar, canine, incisor) and two anomalies (impacted and capped teeth). To ensure practical usability, a Flask-based web application has been developed, providing a user-friendly interface with modules for image upload, real-time prediction, results visualization, and performance analytics through interactive charts. Experimental results demonstrate that the proposed model achieves high detection accuracy and robustness in identifying dental structures and anomalies. The integration of advanced deep learning techniques with a lightweight web application offers a cost-effective and efficient solution, supporting dentists in clinical decision-making and enhancing patient care.