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

Table 3

The output data format of the input layer, every intermediate layer, and the output layer in our final model network.

Layers (type)	Output shape
Conv2d_1 (Conv2D)	(none, 28, 28, 60)
Conv2d_2 (Conv2D)	(none, 24, 24, 60)
Max_pooling2d_1 (MaxPooling1)	(none, 12, 12, 60)
Conv2d_3 (Conv2D)	(none, 10, 10, 30)
Conv2d_4 (Conv2D)	(none, 8, 8, 30)
Max_pooling2d_2 (MaxPooling2)	(none, 4, 4, 30)
Dropout_1 (dropout) (50%)	(none, 4, 4, 30)
Flatten_1 (flatten)	(none, 480)
Dense_1 (dense)	(none, 500)
Dropout_2 (dropout) (50%)	(none, 500)
Output layer	(none, 43)