Table 2.
Maxillary and mandibular absolute mean errors (mm) of landmark points with 95% CI and Cohen’s weighted kappa coefficients of crowding categorization.
Model | ResNet50 | ResNet101 | VGG16 | VGG19 |
---|---|---|---|---|
Maxilla | ||||
Mean error | 0.97 | 0.92 | 0.85 | 0.84 |
95% CI | 0.92–1.02 | 0.87–0.97 | 0.80–0.90 | 0.79–0.89 |
Weighted Kappa | 0.65 | 0.67 | 0.71 | 0.73 |
Mandible | ||||
Mean error | 1.34 | 1.07 | 1.07 | 1.06 |
95% CI | 1.22–1.46 | 0.97–1.17 | 0.97–1.17 | 0.96–1.16 |
Weighted Kappa | 0.61 | 0.68 | 0.71 | 0.73 |
CI confidence interval.