Research Article
Convolutional Neural Network Based Vehicle Classification in Adverse Illuminous Conditions for Intelligent Transportation Systems
Table 3
Class-wise accuracy of the fine-tuned network with different depth layers on our self-constructed dataset.
| Depth | Accuracy (in percentage) | Overall accuracy | Bus | Car | Motorbike | Rickshaw | Truck | Van |
| 18 | 95.89 | 99.68 | 97.77 | 92.82 | 97.37 | 95.56 | 96.52 | 34 | 97.83 | 98.17 | 98.66 | 95.39 | 97.13 | 96.92 | 97.35 | 50 | 97.31 | 98.93 | 97.45 | 94.57 | 97.04 | 98.70 | 97.33 | 101 | 98.51 | 99.02 | 98.34 | 96.51 | 98.93 | 97.27 | 98.10 | 152 | 99.48 | 99.68 | 99.08 | 100 | 100 | 99.65 | 99.65 |
|
|