Table 3. Classification results on VGG19 DCNN features. Best values are shown in bold.
| Classifier | Performance measures | ||||
|---|---|---|---|---|---|
| Sensitivity (%) | Specificity (%) | FNR (%) | Accuracy (%) | F-score | |
| Quadratic SVM | 93.0 | 92.0 | 7 | 92.1 | 0.925 |
| Linear SVM | 88.8 | 88.1 | 11.11 | 88.2 | 0.885 |
| Cubic SVM | 92.0 | 92.0 | 8 | 91.9 | 0.920 |
| Fine KNN | 87.5 | 90.6 | 12.5 | 89.2 | 0.889 |
| Medium KNN | 93.0 | 91.6 | 7 | 90.4 | 0.924 |
| Cubic KNN | 93.0 | 91.6 | 7 | 90.4 | 0.924 |
| Weighted KNN | 91.0 | 91.0 | 9 | 90.9 | 0.910 |
| Subspace discriminant | 90.3 | 90.1 | 9.6 | 90.2 | 0.902 |
| Ensemble subspace KNN | 88.3 | 91.5 | 11.6 | 90.0 | 0.897 |