Table 6.

Average training, validation, and test accuracy achieved by different models through the 10-fold cross-validation.

Model and classifier		AUCa (%)	Sensitivity (%)	Specificity (%)
NasNetLarge
	1st	97.61	90.73	93.42
	2nd	82.15	75.33	81.1
	Combined	93.88	90.15	93.85
Xception
	1st	95.9	88.64	91.78
	2nd	92.25	87.38	81.12
	Combined	83.19	76.09	80.64
InceptionV3
	1st	89.28	83.2	85.51
	2nd	91.69	79.53	81.25
	Combined	89.85	83.29	93.7
InceptionResNetV2
	1st	80.65	74.88	77.92
	2nd	85.41	79.21	81.25
	Combined	85.44	82.69	93.75
ResNet50
	1st	98.73	93.14	95.22
	2nd	94.6	85.64	81.03
	Combined	96.9	93.4	93.72

^aArea under the curve.