Table 3.

Screening performance of tooth extraction diagnosis models.

Model	ResNet50	ResNet101	VGG16	VGG19
Maxilla
Accuracy	0.909	0.915	0.910	0.922
Sensitivity	0.778	0.751	0.811	0.854
Specificity	0.939	0.952	0.933	0.937
AUC	0.961	0.959	0.959	0.961
Mandible
Accuracy	0.895	0.890	0.898	0.898
Sensitivity	0.778	0.716	0.796	0.818
Specificity	0.929	0.941	0.928	0.921
AUC	0.940	0.934	0.945	0.943