Table 9.

Comparison (using CXRs + CT scans): Existing DNNs and the proposed DNN

Metrics	InceptionV3	MobileNet	ResNet	Proposed DNN
Sensitivity (Recall)	0.6935	0.7708	0.9375	0.9792
Specificity	0.744	0.8869	0.878	0.9464
Precision	0.7304	0.8721	0.8848	0.9481
False positive rate	0.256	0.1131	0.122	0.0536
False negative rate	0.3065	0.2292	0.0625	0.0208
Accuracy (%)	71.88	82.89	90.77	96.28
F1 Score	0.7115	0.8183	0.9104	0.9634
AUC	0.8011	0.8951	0.9616	0.9808