Table 5 CNN training parameters and specifications for smart car system.
Parameter | Description |
|---|---|
Network architecture | CNN with custom architecture tailored for smart car system |
Training dataset | - Split: 70% training, 15% validation, 15% testing |
Preprocessing | - Resizing: 224 × 224 pixels |
| Â | - Normalization: Min-Max scaling |
Training algorithm | - Optimization: Adam optimiser |
| Â | - Learning rate: 0.0001 |
| Â | - Loss Function: Categorical Cross entropy |
| Â | - Batch Size: 32 |
Training epochs | - Total: 50 epochs |
|  | - Early Stopping: Patience = 5 epochs |
