COVID-19 Diagnosis Using an Enhanced Inception-ResNetV2 Deep Learning Model in CXR Images

Table 9

The results obtained compared to state-of-the-art methods.

Reference	Utilized models	Highest achievement
Wang et al. [12]	COVID-Net	Accuracy: 92.4% for 2 classes
		83.5% for 4 classes
Hemdan et al. [19]	COVIDX-Net	F1-score: 0.89 for normal
		0.91 for COVID-19
Sethy and Behera [20]	ResNet50 and SVM classifier	Accuracy: 95.38%
Ozturk et al. [29]	Dark COVID-Net	Accuracy: 87.02% for 3 classes
Apostolopoulos and Mpesiana [30]	VGG-19	Accuracy: 93.48% for 3 classes
Khan et al. [18]	CoroNet	Accuracy: 89.6% for 4 classes
		95% for 3 classes
Xu et al. [16]	ResNet	Accuracy: 86.7%
Li et al. [31]	COVNet	Specificity: 96%
		Sensitivity: 90%
		AUC: 96%
Song et al. [15]	DeepPneumonia	Accuracy: 92.4% for 2 classes
Ghoshal and Tucker [32]	Bayesian CNN	Accuracy: 92.90%
Zhang et al. [2]	Deep CNN based on backbone network	Specificity: 70.7%
		Sensitivity: 96.0%
		AUC: 95.2%
Ouchicha et al. [21]	CVDNet	Accuracy: 97.20% for 2 classes
		96.69 for 3 classes (COVID-19 vs. normal vs. viral pneumonia)
Our proposed	Enhanced Inception-ResNetV2	Accuracy: 98.80% (average accuracy) and 99.20 for 3 classes (COVID-19 vs. normal vs. viral pneumonia)
		F1-score: 98.86%
		Precision: 98.61%
		Recall: 99.11
		AUC: 97.2%