Table 2.
Results of the CNNs used for transfer learning
| Network | Accuracy 2-class (%) | Accuracy 3-class (%) | Sensitivity (%) | Specificity (%) |
|---|---|---|---|---|
| VGG19 [10] | 98.75 | 93.48 | 92.85 | 98.75 |
| MobileNet v2 [11] | 97.40 | 92.85 | 99.10 | 97.09 |
| Inception [12] | 86.13 | 92.85 | 12.94 | 99.70* |
| Xception [13] | 85.57 | 92.85 | 0.08 | 99.99* |
| Inception ResNet v2 [12] | 84.38 | 92.85 | 0.01 | 99.83* |
Due to data imbalance, measurements corresponding to Sensitivity and Specificity that obtained not meaningful values for the particular set of results, are denoted by an asterisk (*)