Fig 1.
COVID19 CT scan sample.
Fig 2.
COVID19 chest Xray sample.
Fig 3.
Inception V3 architecture [25].
Fig 4.
VGG19 architecture [28].
Fig 5.
Xception architecture [29].
Fig 6.
ResNet-50 architecture view [27].
Table 1.
Performance comparison among the CNN architectures.
Table 2.
Augmentation types and parameters.
Fig 7.
Comparison of confusion matrix without normalization.
Confusion matrix for InceptionV3 (top left), VGG19 (top right), ResNet-50 (bottom left) and Xception (bottom right).
Fig 8.
Combined ROC curves of Covid CT scan.
Fig 9.
Model accuracy curve for InceptionV3 (top left), VGG19 (top right), ResNet-50 (bottom left) and Xception (bottom right).
Fig 10.
Model loss curve for InceptionV3 (top left), VGG19 (top right), ResNet-50(bottom left) and Xception (bottom right).
Table 3.
Comparison of performance of different models.
Fig 11.
Comparison of confusion matrix without normalization.
Confusion Matrix of InceptionV3 (top left), VGG19 (top right), ResNet-50(bottom left) and Xception(bottom right).
Fig 12.
Combined ROC curves of Covid Xray images.
Table 4.
Comparison of performance of different models.
Fig 13.
Comparison of confusion matrix without normalization.
Confusion Matrix of InceptionV3 (top left), VGG19 (top right), ResNet-50 (bottom left), and Xception (bottom right).
Fig 14.
Combined ROC curves of bacterial X-ray images.
Table 5.
Comparison of performance of different models.
Fig 15.
Comparison of confusion matrix without normalization.
Confusion Matrix of InceptionV3 (top left), VGG19 (top right), ResNet-50 (bottom left) and Xception (bottom right).
Fig 16.
Combined ROC curves of bacterial X-ray images.
Table 6.
Comparison of performance of different models.
Fig 17.
COVID 19 detecting using the Flask app.