Skip to content
master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.

Traffic Sign Classification

Overview

In this project, we will see how deep neural networks and convolutional neural networks are very very powerful to perform image classification tasks. Indeed, we train and validate a model so it can classify traffic sign images using the German Traffic Sign Dataset. After the model is trained, we will then try out your model on images of German traffic signs that we downloaded from the web.

alt text

You will find the code - using Tensorflow - for this project is in the IPython Notebook. More details are available by reading the project notes.

The final model architecture :

Layer Description
Input 32x32x1 Grayscale/ Normalized image
Convolution 5x5 1x1 stride, valid padding, outputs 28x28x6
RELU activation
Max pooling 2x2 stride, valid padding, outputs 14x14x6
Convolution 5x5 1x1 stride, valid padding, outputs 10x10x16
RELU activation
Max pooling 2x2 stride, valid padding, outputs 5x5x16
Fully connected 400 input, 120 output
RELU activation
Dropout 0.9 keep probablility (training)
Fully connected 120 input, 84 output
RELU activation
Dropout 0.9 keep probablility (training)
Fully connected 84 input, 43 output

The final model metrics :

  • Training Set Accuracy: 0.9764310717582703
  • Validation Set Accuracy: 0.9523809552192688
  • Test Set Accuracy: 0.93087885996130659

alt text

About

Build a traffic sign classifier with tensorflow

Resources

Releases

No releases published

Packages

No packages published