Traffic Sign Recognition

Overview

This project implements a Traffic Sign Recognition system using Convolutional Neural Networks (CNN) to classify images of traffic signs. The aim is to build an automated system that accurately recognizes and classifies various traffic signs from images, contributing to the development of advanced driver-assistance systems (ADAS) and autonomous vehicles.

Problem Statement

Traffic signs are critical for ensuring road safety as they convey essential information to drivers. An automated recognition system can help improve safety and efficiency on the roads. For example, a system should recognize a "Stop" sign and alert the driver to stop the vehicle, thereby preventing accidents.

Example Images

Stop Sign:
Yield Sign:
Speed Limit Sign:

Datasets

The dataset used for this project is the German Traffic Sign Recognition Benchmark (GTSRB). It contains over 50,000 images categorized into 43 classes of traffic signs.

Dataset Structure

data/ 
    ├── Train/ # Contains train images organized by class
    ├── Test/ # Contains test images organized by class

CSV Files

Train.csv: Contains paths and labels for the training set.
Test.csv: Contains paths and labels for the test set.

Project Structure

    traffic-sign-recognition/
    │
    ├── data/
    │   ├── Meta/
    │   ├── Test/
    │   ├── Train/
    │   └── ...
    ├── scripts/
    │   ├── data_preprocessing.py
    │   ├── eda.py
    │   ├── model_training.py
    │   ├── evaluation.py
    │   ├── inference.py
    │   └── streamlit_app.py
    ├── main.py
    ├── requirements.txt
    └── README.md

Setup and Installation

Clone the Repository:

    git clone <repository-url>
    cd traffic-sign-recognition
     python3 -m venv env
     source env/bin/activate  # On Windows use `env\Scripts\activate`
     pip install -r requirements.txt

Usage

Data Preprocessing

To preprocess the dataset:

    python main.py --data --data_dir data

Exploratory Data Analysis (EDA)

To generate visualizations and understand the dataset:

python main.py --eda --data_dir data --output_dir outputs

Model Training

To train the CNN model on the preprocessed data:

python main.py --training --data_dir data --model_dir models --epochs 20 --batch_size 64 --learning_rate 0.001

Model Evaluation

To evaluate the trained model:

python main.py --evaluation --data_dir data --model_dir models --output_dir outputs

Inference

To make predictions on new images:

python main.py --inference --model_dir models --image_path path/to/image.jpg

Streamlit Application

To launch the Streamlit application for interactive traffic sign recognition:

streamlit run scripts/streamlit_app.py -- --model_path models/best_model.h5

Notes

Image Paths: Ensure to replace data/Train/00000.png, etc., with actual paths to your images.
Repository URL: Replace <repository-url> with the actual URL of your GitHub repository.
License: Ensure you have a LICENSE file if you include a license section.

This README.md provides a comprehensive overview of your project, making it easy for users to understand its purpose and how to use it effectively. Let me know if you need further adjustments or additions!

deepak2233/Traffic-Signs-Recognition-using-CNN-Keras