Akademik Veri Yönetim Sistemi
JOURNAL OF THE COLLEGE OF PHYSICIANS AND SURGEONS--PAKISTAN : JCPSP, cilt.35, ss.1-7, 2025 (SCI-Expanded, Scopus)
Objective: To assess the performance of a deep learning method for apical lesion segmentation on dental panoramic radiographs.
Study Design: Onservational study.
Place and Duration of the Study: Van Yuzuncu Yil University, Faculty of Dentistry, Istanbul, Turkiye , from March to September 2024.
Methodology: The deep learning model, YOLOv5, based on the YOLO algorithm for apical lesion segmentation was further developed using 1500 anonymised panoramic radiographs. The radiographs have been obtained from the Radiology Archive at University, Department of Oral and Maxillofacial Radiology, Faculty of Dentistry. For apical lesion segmentation, YOLOv5 with the PyTorch model was utilized. The dataset was divided into training (n=1200 radiographs/ 2628 labels), validation (150 radiographs/ 325 labels) and test (n=150 radiographs/ 368 labels) sets. The model's effectiveness was measured using the confusion matrix. Precision, Recall, and F1 Scores provided quantitative assessments of the model's predictive capabilities.
Results: The precision, sensitivity, and F1 score prediction performance values of the YOLOv5 deep learning algorithm are 0.784, 0.682 and 0.729, respectively.
Conclusion: According to this study, periapical lesions on panoramic radiography can be reliably identified using deep learning algorithms. Dental healthcare is being revolutionized by artificial intelligence and deep learning methods, which are advantageous to the system as well as practitioners. While the current YOLO-based system yields encouraging findings, more data should be gathered in future research to improve detection outcomes.
Objective: To assess the performance of a deep learning method for apical lesion segmentation on dental panoramic radiographs.
Study Design: Onservational study.
Place and Duration of the Study: Van Yuzuncu Yil University, Faculty of Dentistry, Istanbul, Turkiye , from March to September 2024.
Methodology: The deep learning model, YOLOv5, based on the YOLO algorithm for apical lesion segmentation was further developed using 1500 anonymised panoramic radiographs. The radiographs have been obtained from the Radiology Archive at University, Department of Oral and Maxillofacial Radiology, Faculty of Dentistry. For apical lesion segmentation, YOLOv5 with the PyTorch model was utilized. The dataset was divided into training (n=1200 radiographs/ 2628 labels), validation (150 radiographs/ 325 labels) and test (n=150 radiographs/ 368 labels) sets. The model's effectiveness was measured using the confusion matrix. Precision, Recall, and F1 Scores provided quantitative assessments of the model's predictive capabilities.
Results: The precision, sensitivity, and F1 score prediction performance values of the YOLOv5 deep learning algorithm are 0.784, 0.682 and 0.729, respectively.
Conclusion: According to this study, periapical lesions on panoramic radiography can be reliably identified using deep learning algorithms. Dental healthcare is being revolutionized by artificial intelligence and deep learning methods, which are advantageous to the system as well as practitioners. While the current YOLO-based system yields encouraging findings, more data should be gathered in future research to improve detection outcomes.