Table 7.
Comparison of the proposed CoroNet with other existing deep learning methods.
| Study | Architecture | Accuracy 3-class (%) | Accuracy 2-class (%) | # Params (in million) |
|---|---|---|---|---|
| Ioannis et al. [43] | VGG19 | 93.48 | 98.75 | 143 |
| Ioannis et al. [43] | Xception | 92.85 | 85.57 | 33 |
| Wang and Wong [42] | Covid-Net (Residual Arch) | NA | 92.4 | 116 |
| Sethy and Behra [45] | ResNet-50 | NA | 95.38 | 36 |
| Hemdan et al. [41] | VGG19 | NA | 90 | 143 |
| Narin et al. [44] | ResNet-50 | NA | 98 | 36 |
| Narin et al. [44] | InceptionV3 | NA | 97 | 26 |
| Ozturk et al. [18] | DarkNet | 87.02 | 98.08 | 1.1 |
| Proposed CoroNet | CoroNet (Xception) | 89.6 | 99 | 33 |