Table 4.

Accuracy, Loss values for the proposed DFL embedded CSDB CNN observed on the training and validation data across 5 folds. Also, the Macro averaged Area under ROC (AUC) and AUC for the COVID-19 class are shown for each fold.

Fold	Accuracy		Loss		Area under ROC (%)
	Training	Validation	Training	Validation	COVID-19	Macro-average
Fold1	99.70	98.23	0.0667	0.167	98.51	97.75
Fold2	99.88	96.79	0.0512	0.237	97.65	97.27
Fold3	99.34	97.61	0.082	0.179	98.26	98.16
Fold4	99.50	98.33	0.088	0.159	98.77	99.09
Fold5	99.81	98.72	0.0710	0.127	98.75	97.75