Table 1 . A summary of recent related studies.

Paper	Dataset	Method	Results
Butt et al. (2020)	219 COVID-19 339 Others	ResNet-18 and ResNet-23	Accuracy = 95.2%
			AUC = 0.996 (99.6%)
			Sensitivity = 98.2%
			Specificity = 92.2%
Li et al. (2020a)	468 COVID-19 2,996 Others	ResNet-50	Accuracy = 89.5%
			AUC = 0.95 (95%)
			Sensitivity = 87%
			Specificity = 92%
Bai et al. (2020)	521 COVID-19 665 others	EfficientNet-B4	Accuracy = 87%
			AUC = 0.9 (90%)
			Sensitivity = 89%
			Specificity = 86%
Amyar, Modzelewski & Ruan (2020)	449 COVID-19 595 others	Two U-Nets	Accuracy = 86%
			AUC = 0.93 (93%)
			Sensitivity = 94%
			Specificity = 79%
Chen et al. (2020)	4382 COVID-19 9,369 others	U-Net++	Accuracy = 95.2%
			Sensitivity = 100%
			Specificity = 93.6
Zheng et al. (2020)	313 COVID-19 229 Others	U-Net and CNN	Accuracy = 90.9%
			AUC = 0.959 (95.9%)
			Sensitivity = 90.7%
			Specificity = 91.1%
Jin et al. (2020a)	496 COVID-19 1,385 Others	Deeplab-v1 and ResNet-152	Accuracy = 94.8%
			AUC = 0.979 (97.9%)
			Sensitivity = 94.1%
			Specificity = 95.5%
Jin et al. (2020b)	723 COVID-19 413 Others	U-Net and ResNet-50	Accuracy = 94.8%
			Sensitivity = 97.4%
			Specificity = 92.2%
Song et al. (2020)	219 COVID-19 399 Others	ResNet-50	Accuracy = 86%
			AUC = 0.95 (95%)
			Sensitivity = 96%
			Precision = 79%
Wang et al. (2020)	325 COVID-19 740 Others	Inception and Adaboosted decision tree	Accuracy = 82.9%
			AUC = 0.9 (90%)
			Sensitivity = 81%
			Specificity = 84%