Artificial Neural Network-Based Deep Learning Model for COVID-19 Patient Detection Using X-Ray Chest Images

Table 6

Comparison of the proposed weight fusion model with other existing deep learning-based studies from the literature.

Method	Target classes	Evaluation results
		Acc.	Prec.	Sens.	Spec.	AUC
Proposed fusion method	3 classes: COVID-19, normal, pneumonia	0.954	0.968	0.991	0.982	0.959
COVID-Net [31]	3 classes: COVID-19, normal, non-COVID-19	0.933	0.989	0.910	—	—
CovidGAN [25]	2 classes: COVID-19, normal	0.950		0.900	0.970	—
Pretrained CNN [17]	2 classes: COVID-19, normal	0.980	1.00	0.960	1.00	—
ResNet18 [18]	5 classes: normal, bacterial, tuberculosis, viral, COVID-19	0.889	0.834	0.859	0.964	—
Triple-view CNN [15]	2 classes: normal, COVID-19	0.998	0.996	0.999	0.997	—
	3 classes: normal, COVID-19, other	0.844
DarkNet [19]	2 classes: COVID-19, no findings	0.980	0.980	0.951	0.953	—
	3 classes: COVID-19, no findings, pneumonia	0.870	0.899	0.853	0.921	—
Deep learning-based decision tree [21]	Multiple classes: COVID-19, TB, non-COVID-19, non-TB	0.950	0.940	0.970	0.930	0.950