Table 5.
Mean overall accuracy after fivefold cross-validation on ‘Test Dataset-2’ using mono-feature CNNs and multi-feature CNNs. Bold denotes the best results obtained
| AlexNet | ResNet50 | SonoNet64 | |
|---|---|---|---|
| CXR(x, y) | 90.59±0.21 | 93.4±0.17 | 91.1±0.8 |
| MF(x, y) | 91.97± 0.24 | 93.17±0.3 | 93.46±0.15 |
| Middle fusion | 92.52±0.32 | 94.26±0.19 | 93.94±0.13 |
| Late-fusion | 92.72±0.17 | 94.44±0.2 | 94.02±0.14 |
| Xception | InceptionV4 | EfficientNetB4 | |
| CXR(x, y) | 92.28±0.46 | 92.99±0.2 | 92.16±0.49 |
| MF(x, y) | 92.61±0.19 | 92.89±0.27 | 93.1±0.17 |
| Middle fusion | 92.89±0.12 | 93.8±0.27 | 93.54±0.29 |
| Late-fusion | 93.77±0.15 | 94.01±0.09 | 93.91±0.07 |