TABLE 4. Comparison With Existing Models.

Set	Precision	Sensitivity	F1 Score	Accuracy
Original CXR images	0.505	0.429	0.46	0.45
Regrouped lung images	0.51	0.45	0.48	0.48
Original CXR images	0.88	0.82	0.85	0.85
AlexNet-SVM	0.736	0.722	0.729	0.74
DenseNet-SVM	0.90	0.76	0.826	0.8219
VGG-SVM	0.852	0.743	0.794	0.7952
PCA-SVM	0.87	0.76	0.81	0.81
Patch-based CNN [6]	0.947	0.90	0.923	0.925
DarkNet and YOLO [7]	0.905	0.835	0.867	0.8702
COVID-SDNet [17]	0.853	0.771	0.81	0.81
Hussain et al. [4]	0.821	0.765	0.792	0.7952
Novitasari et al. [38]	0.96	0.86	0.92	0.9174
Misra, S. et al [39]	0.896	1	0.91	0.933
Jain, R. et al [40]	0.91	0.78	0.86	0.93
Asif Iqbal, et al. [41]	0.90	0.892	–	0.896
Mohammad, et al. [42]	0.85	0.80	–	0.914
Apostolopoulos, et al. [43]	0.933	0.9285	–	0.934
Our method	1	0.88	0.936	0.936