ranuelyn / sleep-detection

A sleep detection project for self-driving cars via Image Processing with AI.

Sleep-Detection

A sleep detection project for self-driving cars via Image Processing with AI.

This Python script, sleep_detection.py, is designed to monitor signs of drowsiness or sleep in individuals, particularly useful for drivers or individuals in critical monitoring jobs. Using real-time video data, the script detects indicators of sleepiness and can trigger alerts to prevent accidents or mishaps.

Features

• Real-time sleepiness detection using facial landmarks.
• Alert system that triggers warnings when drowsiness is detected.
• Ability to process video input from various sources (webcams, recorded video, etc.).
• Adjustable sensitivity to accommodate different user thresholds for drowsiness indicators.

Prerequisites

Before you run the script, ensure that you have the following installed:

• Python 3.x
• OpenCV: For image processing and facial landmark detection.
• Dlib: A toolkit for machine learning and facial landmark detection.
• Playsound: For playing alert sounds when drowsiness is detected.

You can install the required libraries using pip:

pip install opencv-python dlib playsound

Installation

To get started with the sleep detection system, clone this repository:

git clone https://github.com/ranuelyn/Sleep-Detection.git

Usage

To execute the sleep detection script, navigate to the directory where the script is located and run:

python sleep_detection.py

The script will start processing the video feed and monitor for signs of drowsiness in real-time.

Libraries Used

• OpenCV: For image acquisition and processing tasks.
• Dlib: For facial feature detection and machine learning algorithms.
• Playsound: To play an alert sound when the system detects drowsiness.

Contributing

If you would like to contribute to the development of this sleep detection system, please feel free to fork this repository, make your changes, and submit a pull request. Bug reports and feature suggestions can be submitted through the repository's Issues section.