It is a multi-class supervised learning problem, with more than 40 classes and 50k images in total. The data is taken from ‘German traffic Sign Benchmark’ on Kaggle. The model was built using KERAS library (python), and a GUI was built using TKINTER that allowed the users to upload images of traffic sign, and to predict and display its class.
- traffic_sign_imgclass_grayscale(final) : contains the Source-code of the image classifier
- Image Dataset : The German Traffic Sign Benchmark is a multi-class, single-image classification challenge held at the International Joint Conference on Neural Networks (IJCNN) 2011. Single-image, multi-class classification problem More than 40 classes More than 50,000 images in total Large, lifelike database.
dataset link : Dataset
****
Try out the model online
conda create -n traffic python=3.7
activate traffic
pip install -r requirements.txt
1: The template for this GUI is derived from https://github.com/krishnaik06/Deployment-flask