TABLE IV. Comparison of Performances in the Joint Diagnosis and Severity Prediction of COVID-19 With Different Networks on MosMedData.

Network	Diagnostic prediction								Severity prediction
	Normal vs. Mild				Normal vs. Severe				Mild vs. Severe
	Sen. (%)	Spec. (%)	Acc. (%)	F1 (%)	Sen. (%)	Spec. (%)	Acc. (%)	F1 (%)	Sen. (%)	Spec. (%)	Acc. (%)	F1 (%)
VGG-19	54.4 0.37	63.4 0.28	58.4 0.32	58.6 0.33	63.4 0.32	70.8 0.37	65.9 0.28	66.9 0.39	62.7 0.25	65.5 0.23	61.9 0.25	64.1 0.28
ResNet-50	61.1 0.29	65.7 0.33	62.5 0.44	63.3 0.37	66.5 0.26	69.3 0.31	69.1 0.17	67.8 0.33	61.1 0.36	63.8 0.29	64.8 0.34	62.4 0.37
Xception	56.8 0.38	57.9 0.41	59.9 0.27	57.3 0.43	64.8 0.19	67.2 0.26	66.7 0.21	65.9 0.25	62.9 0.33	65.1 0.32	63.1 0.22	63.9 0.34
DenseNet121	59.7 0.27	64.6 0.21	61.1 0.39	62.1 0.28	65.1 0.23	70.1 0.19	67.8 0.29	67.5 0.25	60.2 0.28	64.4 0.21	60.6 0.26	62.2 0.27
InceptionV3	60.4 0.31	62.1 0.38	59.3 0.23	61.2 0.35	66.6 0.28	69.8 0.24	66.2 0.32	68.2 0.31	62.8 0.29	67.9 0.35	61.4 0.19	65.3 0.33
CovTANet (ours)	83.8 0.25	90.3 0.27	85.2 0.19	86.9 0.22	93.9 0.13	96.6 0.11	95.8 0.17	94.2 0.21	90.9 0.16	93.4 0.23	91.7 0.19	92.1 0.11