Drowsy Driver Detection

A real-time fatigue detection application for classifying drowsy and non-drowsy drivers. Capstone project for Computer Vision class.

Table of Contents

• Introduction
• Features
• Requirements
• Installation
• Usage
• How it Works
• License

Introduction

The Fatigue Detection System is designed to monitor the user's alertness by analyzing their eye aspect ratio (EAR) and mouth aspect ratio (MAR) through a webcam. It provides visual and audible alerts when signs of fatigue are detected, thereby reducing the risk of accidents caused by drowsiness.

Features

• Real-time eye tracking to detect fatigue
• Visual indicators showing the user's alertness status
• Audible alarms to alert the user

Requirements

This application requires the following:

• Python 3.6+
• OpenCV
• Mediapipe
• dlib
• PyQt5
• numpy

Installation

Clone the repository to your local machine:

git clone https://github.com/Luchanaaaaa/Fatigue-Detection-System.git
cd Fatigue-Detection-System

Install the required dependencies:

pip install -r requirements.txt

Usage

To run the application, execute the following command:

python main.py

Make sure your webcam is enabled and properly set up before starting the application!

How it Works

• The application uses a webcam to continuously monitor the user.
• Facial landmarks are detected using dlib and Mediapipe's pre-trained model.
• Eye aspect ratios and mouth aspect ratios are calculated to determine the user's level of alertness.
• The system classifies the user's state as active, fatigued, or asleep based on EAR and MAR values and provides corresponding alerts.

License

This project is licensed under the MIT License - see the LICENSE.md file for details.