Research Article
DAFLNet: Dual Asymmetric Feature Learning Network for COVID-19 Disease Diagnosis in X-Rays
Table 4
Best accuracy of nine networks in the test set (%).
| ID | Model | COVID-19 | Normal | Pneumonia | Overall Acc. |
| M1 | EfficientNetV2 | 98.4 | 97.50 | 97.56 | 97.23 | M2 | DenseNet121 | 98.08 | 97.45 | 97.18 | 96.85 | M3 | DenseNet169 | 97.91 | 97.12 | 97.12 | 96.58 | M4 | Eff-V2&D-121 | 98.35 | 97.94 | 98.15 | 97.61 | M5 | Eff-V2&D-169 | 99.13 | 97.77 | 97.77 | 97.34 | M6 | DAFFM-1 | 99.13 | 98.05 | 98.05 | 97.72 | M7 | DAFFM-2 | 99.51 | 98.53 | 98.48 | 98.26 | M8 | DAFLNet-1 | 99.62 | 99.08 | 99.02 | 98.86 | M9 | DAFLNet-2 | 99.62 | 99.13 | 98.86 | 98.81 |
|
|