TABLE 2. ResNet50 Fine-Tuned Architecture.
| Resnet50 |
|---|
| Output from base model |
| Global Average Pooling |
| Fully Connected (64), ReLU |
| Dropout=0.5 |
| Batch Normalization |
| Fully Connected (1), Sigmoid |
| Trainable parameters: 184K |
| Resnet50 |
|---|
| Output from base model |
| Global Average Pooling |
| Fully Connected (64), ReLU |
| Dropout=0.5 |
| Batch Normalization |
| Fully Connected (1), Sigmoid |
| Trainable parameters: 184K |