Research Article
A Lightweight Model for Traffic Sign Classification Based on Enhanced LeNet-5 Network
Table 4
Number of trainable parameters of our proposed network compared with that of previous state-of-the-art approaches.
| Method | Number of trainable parameters (million) |
| Faster R-CNN [41] | 2.92 | Multiscale CNN [42] | 1.4 | Lightweight deep network (student model) [43] | 0.8 | Lightweight deep network (teacher model) [43] | 7.9 | Our first network (our final network) | 0.38 |
|
|