Table 8.

Summary of related multi-class classification for chest X-ray with both pneumonia and COVID-19 conditions.

Study	DL technique	Acc %	Loss function	Optimizer	GPU	Evaluation metric
[42]	CNN	90.64	CCE	RMSprop	–	Acc, P, R, F1-Score, 5-fold cross-validation
[54]	VGG-16	98.3	CCE	Adam	–	Acc, Sn, Sp
[10]	MobileNetV2	94.72	–	Adam	–	Acc, Sn, Sp
[47]	ResNet-50 (COVID-ResNet)	96.23	–	Adam	–	Acc, Sn, P, F1-score
[38]	EfficientNet	96.70	–	Adam	✓	10-fold cross validation, Acc, P, R, F1-score
[67]	VGG-16	95	CCE	Adam	–	Acc, P, R, F1-score
[87]	CapsNet	84.22	MSE	Adam	–	Acc, 10-fold cross validation, Sn, Sp, F1-score, P
[124]	Darknet-19 (DarkCovidNet)	87.02	Cross- entropy	Adam	–	5-fold cross validation, Acc, Sn, Sp, P, F1-score
[91]	SqueezeNet	98.3	–	Bayesian	✓	Acc, COR, COM, Sp, F1-score, MCC
[44]	COVID-Net	93.3	–	Adam	–	Acc, Sn
[103]	COV19-ResNet COV19-CNNet	97.61 94.28	–	–	✓	Acc, P, R, Sp, F1-score,
[76]	ResNet-50	98.18	–	–	✓	Acc, P, R, F1-score
[59]	VGG-CapsNet	92	CCE	SGD	–	Acc, P, R, F1-score, AUROC
[89]	EfficientNet B3-X	93.9	–	Adam	–	Acc, $S{e}_C$, +$P_C$
[106]	CoroNet (Xception)	95	–	Adam	✓	Acc, P, R, Sp, F1-score
[37]	CNN SVM DT KNN	97.14 98.97 96.10 95.76	–	Adam, Bayesian	✓	5-fold cross validation, Acc, Sn, Sp, AUROC, F1-score
[111]	CNN (CVDNet)	96.69	cross-entropy	Adam	–	Acc, P, R, F1-score, 5-fold cross validation
[58]	Ensemble (ResNet-50V2, VGG-16, InceptionV3)	95.49	CCE	Adam	✓	Acc, Sn, Sp, P, AUROC
[79]	ResNet-50	92	CCE	Adam	–	Acc, Sn, Sp, F1-score, AUROC
[60]	SqueezeNet & MobileNetV2 (Combined features set)	99.27	–	–	✓	Acc, Sn, Sp, P, F1-score, 5-fold cross validation