AI Model Identifies Twelve Dental Implant Brands with 96.23% Accuracy on Radiographs

A two-stage deep learning system (YOLOv11 detection then EfficientNetV2-M classification) identified twelve implant brands on panoramic and periapical radiographs with 96.23% accuracy (95% CI 94.5–97.7%) and an end-to-end precision-aware F1-score of 88.2%.

• On a held-out test set of 531 implant crops, macro-averaged F1 was 0.953 and AUC 0.991.
• End-to-end pipeline, counting false detections, yielded precision 81.6%, recall 96.0%, and F1 88.2%.
• Grad-CAM heatmaps confirmed the model focused on clinically relevant implant morphology.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.