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

Table 8

Model classification results for 3 categories (COVID-19 vs. normal vs. viral pneumonia) classification task for the new release of the dataset.

Model	Category	Accuracy (%)	F1-score (%)	Precision (%)	Recall (%)
Inception-ResNetV2	COVID-19	98.02	96.00	98.36	93.75
	Normal	97.36	97.97	97.15	98.80
	Viral pneumonia	98.94	95.09	94.74	95.46
Xception	COVID-19	44.85	36.20	15.57	21.77
	Normal	50.73	64.16	63.53	64.80
	Viral pneumonia	89.38	24.24	5.14	8.48
VGG16	COVID-19	99.27	97.66	98.43	98.04
	Normal	98.81	99.00	98.61	98.80
	Viral pneumonia	99.54	94.70	95.42	95.06
ResNet50V2	COVID-19	57.92	52.87	39.42	45.16
	Normal	52.24	55.70	74.27	63.66
	Viral pneumonia	91.16	61.36	32.27	42.30
InceptionV3	COVID-19	98.88	98.70	97.93	98.31
	Normal	98.35	98.70	98.90	98.80
	Viral pneumonia	99.21	95.46	96.18	95.82
MobileNetV2	COVID-19	98.68	98.96	96.94	97.94
	Normal	98.15	98.30	99.09	98.70
	Viral pneumonia	98.15	94.70	94.70	94.70
DenseNet121	COVID-19	99.34	98.96	99.22	99.09
	Normal	98.48	98.90	99.40	99.15
	Viral pneumonia	99.14	98.49	94.20	96.30
ResNet101V2	COVID-19	99.21	98.18	98.95	98.56
	Normal	98.68	98.90	99.00	98.95
	Viral pneumonia	99.47	97.73	94.85	96.27