Traffic Sign Recognition

Implementation of a Convolutional Neural Network based on the ResNet architecture, trained for the recognition of traffic signs with the GTSRB dataset.

Repository Structure

• utils.py: contains helper functions
• preprocessing.py; functions for image processing
• loader.py: loads the images and labels into the appropriate format
• training.py: CNN training
• _architecture.png: diagram of the implemented network

How to Run

1. 
2. That's it

OR

1. Download the training dataset here
2. Clone this repo to your local machine and extract the downloaded zip into the same directory. Make sure you have the modules required installed (pip install - r requirements.txt)
3. Run loader.py
4. Run training.py