Table 4 Experimental Parameters for Proposed Model.
Parameters | Details |
|---|---|
Batch size | 8 |
Data augmentation strategies | Flipping, zooming, translations and rotations, |
Normalization | 0, 1 |
Regularization | L2 regularization (Weight decay) |
Optimizer | ‘Adam optimizer’ |
Dropout rate | 0.1 |
Epochs | 100 |
Image input size | (224 × 224) |
Hyperparameter optimization | Bayesian optimization |
Transfer learning | MobileNet-V3 |
Loss function | Multi-class categorical cross-entropy function |
Learning rate | 0.2 |
Growth rate | 24 |
Split ratio | Training: testing: validation 70: 15: 15 |
Shuffling in database | “YES” |
Brightness range | [0.2, 2.25] |
Rotation | 0 to 15 Degrees |
