Table 5.

Comparison of test result for pre-trained and ensemble model for the alternative dataset in one holdout run. Bold means it outperformed other models.

Model	Accuracy	Precision	Sensitivity	F1	Specificity
ResNet50V2	0.9565	0.9710	0.9436	0.9571	0.9701
ResNet152V2	0.9510	0.9347	0.9347	0.9502	0.9673
MobileNet	0.9710	0.9420	1.0000	0.9701	0.9452
F-EDNC	0.9783	0.9565	1.0000	0.9778	0.9583
O-EDNC	0.9710	0.9420	1.0000	0.9701	0.9452
FC-EDNC	0.9783	0.9714	0.9602	0.9734	0.9602
CANet	0.9237	0.9168	0.9205	0.9223	0.9115