The paper describes the design of an efficient sleepy driver detection algorithm utilizing the Raspberry Pi CPU and the "facial 68 landmark" model for accurate facial recognition. The system creates a threshold to detect fatigue by analyzing the eye aspect ratio (EAR); An EAR of less than 0.25 indicates fatigue and an EAR of less than 0.15 indicates sleepiness. Various tests with 10 volunteers showed good performance, achieving 97% accuracy by agreement, even with modifications such as wearing glasses. The system maintains 96% accuracy across 400 test frames and different conditions, including different lenses and images. The system also integrates MQ3 sensors for alcohol detection and vibration, and ultrasonic sensors for collisions, improving its performance and safety. Additionally, if alcohol is detected or an accident occurs, the GPS module will instantly send location information. These results demonstrate the potential for practical use of this system to improve driver safety and health.
No takes yet. Share an insight, caveat, or question.
Kumar et al. (2024) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: