Table 2.

Comparison of evaluation indicators for different classifiers in the test set.

Model	CNN	DenseNet121	ResNet50V2	CNN	DenseNet121	ResNet50V2
Method	Supervised Learning	Supervised Learning	Supervised Learning	Supervised Contrastive Learning	Supervised Contrastive Learning	Supervised Contrastive Learning
Accuracy	0.943	0.934	0.957	0.959	0.960	0.961
Sensitivity	0.322	0.230	0.533	0.596	0.564	0.599
Specificity	0.998	0.996	0.995	0.991	0.995	0.993
Prevalence	0.081	0.081	0.081	0.081	0.081	0.081
Precision	0.930	0.840	0.900	0.858	0.908	0.878
NPV	0.943	0.936	0.960	0.965	0.963	0.965
F1 Score	0.479	0.361	0.669	0.704	0.696	0.712
TP	147	105	243	272	257	273
FP	11	20	27	45	26	38
FN	309	351	213	184	199	183
TN	5148	5139	5132	5114	5133	5121