A Lightweight Model for Traffic Sign Classification Based on Enhanced LeNet-5 Network

Table 6

Accuracy of our proposed network compared with that of previous state-of-the-art approaches on German traffic sign data set.

Methods	Accuracy (%)
Optimized LeNet-5 [34]	Between 93 and 96
Modified LeNet-5 network [35]	95.2
Improved LeNet-5 [27]	99.75
LeNet architecture [45]	96.23
Small CNN [46]	97.4
Improved LeNet-5 [44]	98.12
Lightweight deep network [43]	99.61
Deep CNN [37]	98.96
Efficient CNN [41]	99.66
Ours	99.84