Table 7.

Performance comparison of eight models on test set

	ACC	AUC	$F_1$	Rec	Prec
CovidVGG${}^{\mathrm{a}}$	0.941	0.987	0.926	0.918	0.935
VGG19	0.960	0.994	0.946	0.934	0.959
Resnet50	0.964	0.996	0.953	0.945	0.961
AlexNet	0.937	0.982	0.912	0.913	0.910
Single-CNN${}^{\mathrm{a}}$	0.951	0.986	0.932	0.931	0.932
Double-CNN${}^{\mathrm{a}}$	0.959	0.989	0.948	0.944	0.953
TIN	0.974	0.998	0.958	0.962	0.954
DNN	0.957	0.990	0.938	0.922	0.955
OriginalViT	0.956	0.997	0.945	0.958	0.935
CovidViT${}^{\mathrm{a}}$	0.980	0.998	0.974	0.978	0.969

The best results are in bold

${}^{\mathrm{a}}$Represents our proposed models