Table 4 Comparisons between the studies in the literature on germ detection and this study
From: Automatic detection of developmental stages of molar teeth with deep learning
Author | Task | Type of Image | Model | Number of Classes | Data size | Metrics |
|---|---|---|---|---|---|---|
Çalışkan et. al | Object Detection | Panoramic | Faster R-CNN | 1 | 74 | Accuracy: 0.8372, Sensitivity: 0.4545 Specificity: 0.9688 Precision: 0.8333 |
De Tobel et. al | Classification | Panoramic | AlexNet | 10 | 400 | Mean accuracy: 0.51, Mean absolute difference: 0.6, Mean linearly weighted kappa: 0.82 |
Merdietio et. al | Classification | Panoramic | DenseNet201 | 10 | 400 | Accuracy: 0.61 mean absolute difference: 0.53 linear Cohen’s kappa coefficient: 0.84 |
Banar et. al | Segmentation | Panoramic | U-Net like CNN model | 10 | 400 | Dice score: 93% Accuracy: 54% mean absolute error: 0.69 linear Cohen’s kappa coefficient: 0.79 |
Kaya et. al | Object Detection | Panoramic | YOLOv4 | 1 | 4518 | Average Precision: 94.16% Precision: 0.89 Recall: 0.91 F1-score: 0.90 |
This work | Object Detection | Panoramic | Cascade R-CNN YOLOv3 HTC DetectoRS SSD EfficientNet NAS-FPN Deformable DETR PAA | 4 | 210 | Avg. Accuracy: 0.81 Average Precision: 0.86 Average Recall: 0.85 Average F1-score: 0.86 |