Table 11.

Performance comparison of the proposed Covid-19 diagnostic method with other deep learning methods

Labels	Researcher	Model	Accuracy (%)
Binary labels	Wehbe et al. [22]	DeepCOVID-XR	83
	Sethy et al. [24]	ResNet50+SVM	95.38
	Loey et al. [26]	ResNet50 with augmentation	82.91
	Kawsher Mahbub et al. [27]	DNN	99.87
	Mukherjee et al. [28]	DNN	96.28
	Das et al. [33]	TIN	98.77
Multiple labels	Ozturk [6]	DarkCovidNet	87.02
	Apostolopoulos et al. [7]	VGG-19	93.48
	Al-Falluji [8]	Modified ResNet18-Based	96.73
	Wang et al. [23]	COVID-Net	92.4
	Dev et al. [25]	HCN-DML	96.67
	Das et al. [33]	TIN	97.4${}^{\mathrm{a}}$
	Kawsher Mahbub et al. [27]	DNN	95.7${}^{\mathrm{a}}$
	Proposed study	CovidViT	98.0

${}^{\mathrm{a}}$Denotes the accuracy is gained by training on the same dataset with CovidViT by ourselves