Table 2.
Performance of VGG16, MobileNetV2, ResNet18, and AlexNet on training and testing set
| Models | Splits | Classes | Precision | Recall | F1-Score | Accuracy |
|---|---|---|---|---|---|---|
| VGG16 | Training | Normal | 0.97 | 0.99 | 0.98 | 0.98 |
| Viral Pneumonia | 0.99 | 0.97 | 0.98 | |||
| COVID-19 | 0.99 | 0.99 | 0.99 | |||
| Testing | Normal | 0.91 | 0.93 | 0.91 | 0.97 | |
| Viral Pneumonia | 0.91 | 0.9 | 0.904 | |||
| COVID-19 | 0.93 | 0.95 | 0.939 | |||
| AlexNet | Training | Normal | 0.96 | 0.99 | 0.98 | 0.98 |
| Viral Pneumonia | 0.98 | 0.96 | 0.96 | |||
| COVID-19 | 0.98 | 1 | 0.99 | |||
| Testing | Normal | 0.94 | 0.92 | 0.92 | 0.98 | |
| Viral Pneumonia | 0.98 | 0.93 | 0.94 | |||
| COVID-19 | 0.97 | 0.98 | 0.979 | |||
| MobileNetV2 | Training | Normal | 0.94 | 1 | 0.97 | 0.97 |
| Viral Pneumonia | 1 | 0.93 | 0.96 | |||
| COVID-19 | 0.98 | 1 | 0.99 | |||
| Testing | Normal | 0.91 | 0.95 | 0.92 | 0.97 | |
| Viral Pneumonia | 0.94 | 0.97 | 0.95 | |||
| COVID-19 | 0.90 | 0.93 | 0.91 | |||
| ResNet18 | Training | Normal | 0.95 | 0.98 | 0.96 | 0.96 |
| Viral Pneumonia | 0.98 | 0.95 | 0.96 | |||
| COVID-19 | 1 | 0.94 | 0.97 | |||
| Testing | Normal | 0.88 | 0.96 | 0.91 | 0.89 | |
| Viral Pneumonia | 0.82 | 0.93 | 0.87 | |||
| COVID-19 | 0.92 | 0.73 | 0.81 |