Table 2.
Performance (accuracy) of the different configurations for data augmentation.
| DataAUG | VIR | BARK | GRAV | POR |
|---|---|---|---|---|
| NoDA | 85.53 | 87.48 | 97.66 | 86.29 |
| App1 | 87.00 | 89.60 | 97.83 | 87.05 |
| App2 | 86.87 | 90.17 | 98.08 | 85.97 |
| App3 | 87.80 | 89.45 | 97.99 | 87.05 |
| App4 | 86.33 | 87.91 | 97.74 | 84.90 |
| App5 | 86.00 | 87.61 | 97.83 | 86.41 |
| App6 | -- | 88.63 | 98.08 | 87.37 |
| App7 | -- | 89.28 | 97.99 | 88.13 |
| App8 | -- | 87.29 | 97.74 | 86.06 |
| App9 | 85.67 | 88.86 | 98.24 | 86.19 |
| App10 | 84.20 | 86.39 | 98.41 | 85.10 |
| App11 | 85.47 | 89.20 | 97.91 | 86.71 |
| [29] | 82.93 | -- | -- | -- |
| [33] | 83.07 | -- | -- | -- |
| EnsDA_all | 90.00 | 91.27 | 98.33 | 89.21 |
| EnsDA_5 | 89.60 | 91.01 | 98.08 | 88.56 |
| EnsBase | 89.73 | 90.67 | 98.16 | 87.58 |
| EnsBase_5 | 89.60 | 90.66 | 97.99 | 87.48 |
| State of the art | 89.60 | 90.40 | 98.21 | 80.09/90.08 * |
* As noted above, for fair comparison, 80.09 is the best performance using their deep learning approach, but 90.08 was obtained when combining handcrafted with deep learning features. Note: the virus data set has gray level images; for this reason, the three data augmentation methods based on color (App7–8) perform poorly on VIR, so these methods are not reported for this data set. Additionally, because of the low performance on VIR, [29,33] are not tested on BARK, GRAV, and POR. Bold values highlight the best results.