Table 5.
Results obtained on ensembling various combinations of base learners on all the three datasets used in this study.
| Model-1 | Model-2 | Model-3 | Ensemble result (classification accuracy %) | ||
|---|---|---|---|---|---|
| Mendeley LBC | SIPaKMeD 2-Class | SIPaKMeD 5-Class | |||
| Inception v3 | Xception | DenseNet-121 | 96.05 | 95.38 | 92.30 |
| Inception v3 | Xception | DenseNet-201 | 94.04 | 93.89 | 90.60 |
| Inception v3 | VGG-16 | DenseNet-169 | 97.37 | 96.39 | 93.01 |
| Xception | VGG-16 | ResNet-50 | 95.06 | 93.98 | 91.05 |
| DenseNet169 | VGG-19 | ResNet-50 | 96.36 | 94.68 | 91.56 |
| DenseNet169 | VGG-19 | ResNet-101 | 95.64 | 93.07 | 90.42 |
| Inception v3 | Xception | DenseNet-169 | 99.23 | 98.55 | 95.43 |