Table 3 Performance (in %) comparison of deep learning models on the proposed datasets.
Dataset | Model | Accuracy | Precision | Recall | F1-Score |
|---|---|---|---|---|---|
Kaggle dataset | CNN | 86.85 | 80.56 | 81.25 | 83.06 |
VGG16 | 84.75 | 85.00 | 85.00 | 85.00 | |
MobileNetV2 | 82.03 | 84.00 | 79.00 | 81.00 | |
XceptionNet | 92.97 | 90.25 | 89.56 | 91.36 | |
Proposed MADRN | 94.78 | 91.00 | 93.00 | 92.00 | |
Blended dataset | CNN | 83.23 | 85.00 | 83.00 | 83.00 |
VGG16 | 85.51 | 86.00 | 85.00 | 86.00 | |
MobileNetV2 | 84.88 | 85.00 | 85.00 | 85.00 | |
XceptionNet | 79.67 | 79.00 | 80.00 | 79.00 | |
Proposed MADRN | 92.25 | 93.00 | 92.00 | 92.00 |
