Multi-Index Driver Drowsiness Detection Method Based on Driver's Facial Recognition Using Haar Features and...

Area of Science:

• Road safety
• Human-computer interaction
• Biomedical engineering

Background:

• Road accidents linked to driver fatigue are severe, necessitating advanced detection systems.
• Current drowsiness detection methods vary, with driver monitoring technologies underexplored.
• Existing systems may rely on steering or vehicle data, but driver-focused monitoring offers potential.

Purpose of the Study:

• To evaluate a multidimensional drowsiness index using facial expressions, gaze, and head position.
• To assess the feasibility of implementing this system in a low-cost electronic package.
• To develop and compare algorithms for real-time driver drowsiness detection.

Main Methods:

• Developed a drowsiness detection algorithm using facial features (blinking, yawning, eye-opening), gaze direction, and head position.
• Compared Haar features and Histograms of Oriented Gradients (HOG) for facial recognition.
• Implemented the system on a Raspberry Pi for low-cost prototyping.

Main Results:

• The multidimensional index demonstrated superior performance in detecting driver drowsiness compared to single-index approaches.
• Facial recognition algorithms based on Haar features and HOG were successfully implemented.
• A functional, low-cost prototype for drowsiness detection was created.

Conclusions:

• A multidimensional approach to driver drowsiness detection is more effective than single-parameter methods.
• Low-cost hardware like Raspberry Pi is suitable for developing practical driver monitoring systems.
• This technology has the potential to significantly enhance road safety by mitigating fatigue-related accidents.