Traffic Sign Detector Using Keras ⛔️

A deep learning project using Keras to detect traffic signs, specifically focusing on stop signs and traffic lights.

---

(GUI Interface with Image Selection)

Features

• Real-time traffic sign detection
• Modern GUI interface
• Support for stop signs, traffic lights, and other traffic signs
• High accuracy predictions
• Easy-to-use interface

Project Structure

• src/: Source code for the project
  • train.py: Script for training the model
  • predict.py: Script for making predictions
  • gui.py: Modern GUI application
  • utils.py: Utility functions
• model/: Directory for storing the trained model
• data/: Directory for storing the dataset
  • train/: Training data
  • test/: Test data
• requirements.txt: List of Python dependencies
• README.md: Project documentation

Prerequisites

• Python 3.8 or higher
• pip package manager
• Virtual environment (recommended)

Getting Started

1. Clone the repository:

git clone https://github.com/byrongomezjr/traffic_sign_detection.git
cd traffic_sign_detection

2. Create and activate virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install the required dependencies:

pip install -r requirements.txt

Dataset Setup

This project uses the GTSRB - German Traffic Sign Recognition Benchmark dataset.

1. Download the dataset:

   • Visit GTSRB Dataset on Kaggle
   • Download archive.zip

2. Prepare the dataset:

   • Create a temp_data directory in the project root
   • Place the downloaded archive.zip in temp_data/
   • Run: python organize_dataset.py

Usage

1. Train the model:

python src/train.py

2. Make predictions using the GUI:

python src/gui.py

Model Performance

• The model is trained specifically on three classes:
  • Stop signs
  • Traffic lights
  • Other traffic signs
• Achieves high accuracy on test data
• Fast prediction times

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

MIT License

Author

Byron Gomez Jr.

Acknowledgments

• GTSRB Dataset
• Keras Documentation