Table 7.

Diagnosis results of the DenseANet on the COVID-CT-Dataset. The numbers in bold black font represent the best classification results.

Models	Acc	AUC	Se	Sp	Pr	F1
ResNet101	0.8633	0.9274	0.8797	0.8510	0.8165	0.8469
ResNet101+ARL	0.8793	0.9416	0.8968	0.8661	0.8347	0.8646
Baseline	0.8903	0.9473	0.9083	0.8769	0.8476	0.8769
Baseline + ARL	0.8941	0.9607	0.9112	0.8812	0.8525	0.8809
COVID-AL	0.8670	0.9318	0.8739	0.8618	0.8266	0.8496
COVID-CT-DenseNet	0.8879	0.9475	0.9026	0.8769	0.8468	0.8738
COVIDNet-CT	0.8916	0.9514	0.9083	0.8791	0.8499	0.8781
DenseANet	0.9027	0.9564	0.9226	0.8877	0.8610	0.8907