TABLE 2. Recent AI Research for COVID-19.

Reference	Methods	# of samples	# of classes	Type of Images	Sensitivity	Specificity	Accuracy	AUC
X. Wu et al. (2020) [41]	Multi-view deep learning model (ResNet50 based)	495	2	CT images	81.1%	61.5%	76%	0.819
A. A. Ardakani et al. (2020) [42]	Deep learning technique (ResNet-101 based)	1020	2	CT images	100%,	99.02%,	99.51%	0.994
	Deep learning technique (Xception based)	1020	2	CT images	98.04%,	100%	99.02%	0.994
K. Zhang et al. (2020) [43]	AI system (ResNet-18 based)	3,777	3	CT images	94.93%	91.13%	92.49%	0.981
H. Panwar et al. (2020) [44]	nCOVnet, transfer learning, deep CNN	337	2	X-ray images	97.62%	89.13%	88.10%	0.881
C. Butt et al. (2020) [45]	Multiple CNN models (ResNet-18 based)	618	3	CT images	98.2%	92.2%	86.7%	0.996
M. Nour et al. (2020) [46]	Training CNN model, deep feature extraction, SVM	2,905	3	X-ray images	89.39%	99.75%	98.97%,	0.994
X. Wang et al. (2020) [47]	Weakly supervised deep learning framework	450	3	CT images	94.5%	95.3%	96.2%	0.970