Disclaimer: The model trained on the provided traffic sign images does really well on the test set images, but does poorly on real life images consisting of traffic signs in NYC and Europe. It looks like the model is not "smart" enough to generalize well. Please see more discussion in Traffic_Sign_Classifier-tz.ipynb.

Deep Learning

Project: Build a Traffic Sign Recognition Program

Overview

In this project, you will use what you've learned about deep neural networks and convolutional neural networks to classify traffic signs. You will train a model so it can decode traffic signs from natural images by using the German Traffic Sign Dataset. After the model is trained, you will then test your model program on new images of traffic signs you find on the web, or, if you're feeling adventurous pictures of traffic signs you find locally!

Dependencies

This project requires Python 3.5 and the following Python libraries installed:

• Jupyter
• NumPy
• SciPy
• scikit-learn
• TensorFlow
• Matplotlib
• Pandas (Optional)

Run this command at the terminal prompt to install OpenCV. Useful for image processing:

• conda install -c https://conda.anaconda.org/menpo opencv3

Dataset

1. Download the dataset. This is a pickled dataset in which we've already resized the images to 32x32.
2. Clone the project and start the notebook.

git clone https://github.com/udacity/CarND-Traffic-Signs
cd CarND-Traffic-Signs
jupyter notebook Traffic_Signs_Recognition.ipynb

3. Follow the instructions in the Traffic_Signs_Recognition.ipynb notebook.