Table 2.
Slice-level results of different methods on COVID-19 Dataset-A
| Methods | Acc (%) | Sen (%) | Spe (%) | AUC(%) | p |
|---|---|---|---|---|---|
| ResNet [43] | 95.11 | 96.19 | 93.70 | 99.28 | 4.4e-8 |
| ResNet+CBAM [40] | 95.45 | 95.79 | 95.01 | 99.46 | 3.7e-14 |
| SE-Net [39] | 96.02 | 98.00 | 93.44 | 99.53 | 7.3e-8 |
| Ma et al. [29] | 97.12 | 97.21 | 96.89 | 99.34 | 1.9e-5 |
| Ours (ResNet50) | 98.18 | 98.60 | 97.64 | 99.84 |
Acc: accuracy, Spe: specificity, Sen: sensitivity (best in bold)