We examine the use of video data to determine a driver's drowsiness level. We conduct a user study to collect video of a user reading, watching a driving simulation, and playing a video game that simulates driving. Alongside each video is the user's sleepiness as measured by the Stanford Sleepiness Scale, the Epworth Sleepiness Scale, and eight questions that have been shown to coincide with unsafe driving. Using this data, we replicate the results of prior art, showing that on average, changes in eye movement do correlate with drowsiness. We find however, that the measurements appear to have no predictive value for the drowsiness metrics that are known to coincide with unsafe driving. We determine that additional research in detecting driver drowsiness is needed. Our user study data is publicly available.
On Using Drivers' Eyes to Predict Accident-Causing Drowsiness Levels
2018-11-01
925522 byte
Conference paper
Electronic Resource
English
Drowsiness Detection for Drivers using IoT
IEEE | 2023
|Drowsiness Detector Using Python For Reducing Accident
IEEE | 2022
|Drivers drowsiness detection in embedded system
IEEE | 2007
|Automatic Drowsiness Detection to Minimize Road Accident
Springer Verlag | 2025
|Monitoring drivers drowsiness using a wide angle lens
IEEE | 2013
|