Table 6 False negative rate (FNR) results (mean and standard deviation) of the five-fold cross-validation evaluated on the consensus test set. The best metrics are highlighted in bold
From: AI-Dentify: deep learning for proximal caries detection on bitewing x-ray - HUNT4 Oral Health Study
Model | Enamel caries | Dentine caries | Secondary lesion | mFNR |
|---|---|---|---|---|
YOLOv5 M | 0.153 ± 0.021 | 0.215 ± 0.032 | 0.160 ± 0.035 | 0.176 ± 0.025 |
RetinaNet | 0.185 ± 0.060 | 0.364 ± 0.044 | 0.200 ± 0.045 | 0.250 ± 0.031 |
EfficientDet D0 | 0.606 ± 0.060 | 0.636 ± 0.020 | 0.533 ± 0.020 | 0.592 ± 0.025 |
EfficientDet D1 | 0.479 ± 0.018 | 0.524 ± 0.020 | 0.459 ± 0.021 | 0.487 ± 0.011 |