Driver Drowsiness Detection System

A real-time driver drowsiness detection system using eye state analysis. Captures driver images via mobile camera and uses machine learning to detect signs of drowsiness.

Table of Contents

• Project Overview
• System Architecture
• Features
• Installation
• Usage
• Alert Notification
• Contributing
• Contact
• Demo

Project Overview

This system consists of two main components:

1. Mobile Application: Captures driver images (2 images/minute) and communicates with API
2. Detection Model: Uses Haar cascade classifier for face detection and pre-trained CNN for eye state classification

Workflow:

1. Mobile app captures and sends images to API
2. Server processes images using OpenCV and ML model
3. Returns eye state classification with confidence score
4. Triggers alerts when eyes are closed beyond threshold

System Architecture

Features

• Real-time image capture and processing
• Haar cascade face detection
• Eye state classification (Open/Closed)
• Confidence percentage for predictions
• REST API integration
• Customizable alert thresholds
• Multi-platform support

Installation

Prerequisites

• Python 3.10+
• OpenCV
• TensorFlow/Keras (for model loading)
• Flask (for API server)
• Android/iOS development environment (for mobile app)

Setup

1. Clone repository:

git clone https://github.com/Wahid234/driver-drowsiness-detection.git
cd driver-drowsiness-detection

2. Install Python dependencies:

pip install -r requirements.txt

Usage

Mobile Application

1. Place phone in vehicle facing driver
2. Launch application
3. Grant camera permissions
4. System will auto-capture images and send to API

Backend Server

1. Detection Model
2. API Integration

Customizable alert Notification types:

1. 🔊 Audio alarm
2. 📳 Vibration

🤝 Contributors

• Wahid Alzubeir (Detection Model and API)
• Essa Shehab (Flutter)

Contact

For questions or support, contact: Wahid Alzubeir - [email protected]

Project Link: https://github.com/Wahid234/driver-drowsiness-detection

Demo