We are pleased to announce that our latest research has been published in BMC Oral Health (Springer Nature, Open Access).
The study, titled “Teeth identification and numbering in mixed dentition: evaluating deep learning models for pediatric panoramic radiographs,” compares state-of-the-art detection and segmentation architectures — YOLOv8, YOLOv11, Mask R-CNN and DeepLabV3 — for automated tooth detection and FDI numbering in children aged 6 to 12 years. A total of 1,378 anonymized panoramic radiographs were annotated with polygon labels and stratified by age to assess age-specific performance.
YOLOv11 achieved the highest scores across all metrics (Precision 0.8435, Recall 0.8755, F1 0.8592, mAP50 0.8715) and consistently outperformed YOLOv8 in the age-based analysis, reaching an F1 score of 0.9657 and mAP50 of 0.9817 at age 12, where dentition is more stable. The findings support YOLO-based models as promising decision-support tools for standardized charting during the mixed dentition period.
Authors:
Esra Özçelik (Ordu University)
Hüseyin Şimşek (Ordu University)
Abdulsamet Aktaş (Marmara University)
Hamza Osman İlhan (Yıldız Technical University)
Yasin Yaşa (Ordu University)
