Table 1. Multi-class image classification performances of CNN algorithms on the test sets compared with the average performance of clinicians; ‘oral and maxillofacial surgeons’ vs. ‘GPs’.

	Class
	OSCC					OPMDs
	Precision	Recall (Sensitivity)	Specificity	F1 score	AUC of ROC curve	Precision	Recall (Sensitivity)	Specificity	F1 score	AUC of ROC curve
DenseNet-169	0.98	0.99	0.99	0.98	1.0	0.95	0.95	0.97	0.95	0.98
ResNet-101	0.96	0.92	0.94	0.94	0.99	0.97	0.97	0.94	0.97	0.97
SqueezeNet	0.85	0.72	0.92	0.78	0.88	0.76	0.78	0.88	0.77	0.87
Swin-S	0.69	0.73	0.83	0.71	0.71	0.63	0.74	0.88	0.68	0.80
Oral and maxillofacial surgeons	-	0.90	0.89	-	-	-	0.74	0.93	-	-
GPs	-	0.77	0.87	-	-	-	0.68	0.86	-	-

AUC, area under the curve; ROC, receiver operating characteristics; GPs, General practictioners.