Traffic Sign Recognition

Repository Outline

1. Main.ipynb - Main jupyter notebook used for modeling.
2. Inference.ipynb - Notebook for inference trained model.
3. requirements.txt - text file listing all requirements to run the notebook
4. url.txt - URL of the deployment, and dataset

Problem Background

Recognizing traffic signs accurately is essential for developing safe and efficient autonomous driving systems. Misinterpreting signs can lead to accidents or traffic violations. The goal of this project is to build a reliable model that can classify traffic signs from images, and further on, can be used in self-driving technologies, road monitoring, and driver assistance systems.

Project Output

Result of this project is a model that can predict a picture of a traffic sign and classify them into their group.
The project is deployed on streamlit for user convenience.

Data

The model was trained on the dataset taken from kaggle, Traffic Sign Dataset - Classification

Method

This app uses a Convolutional Neural Network (CNN) model trained to classify various traffic signs. The model used Transfer Learning from MobileNetV2

The model achieved:

• F1 Score on Test Set: 0.95
• F1 Score on Training Set: 0.98

Reference

Streamlit Deployment link.
Main.ipynb for the notebook. please install the required library by running pip install -r requirements.txt