Table 5.
Accuracy of all models. ACC stands for accuracy. AD stands for augmented data; this indicates that the model is trained with augmented data. MD stands for metadata, which indicates that the model is trained with metadata. The bold numbers highlight the highest performance. These results are calculated from the confusion matrix, the two highest model confusion matrices are DenseNet201 and InceptionResNetV2 (as shown in Figure 10 and Figure 11).
| Model | ACC (AD) | ACC (MD) |
|---|---|---|
| InceptionResNetV2 | 0.79 | 0.90 |
| DenseNet201 | 0.84 | 0.89 |
| ResNet50 | 0.76 | 0.70 |
| ResNet152 | 0.81 | 0.57 |
| NasNetLarge | 0.56 | 0.84 |
| MobileNetV2 | 0.83 | 0.81 |
| MobileNetV3Small | 0.83 | 0.78 |
| MobileNetV3Large | 0.85 | 0.86 |
| NasNetMobile | 0.84 | 0.86 |