Robust Real-time Driver Drowsiness Detection System for Heterogeneous Lightning Conditions
Abstract
Driver Fatigue Detection (DFW) has become the most active research area in the field of Advanced Driver Assistance Systems (ADAS). Its robust performance under real driving conditions requires the use of Infra-Red illumination. Direct IR illumination systems represent the current option for manufacturers and OEM, according to comfort and ergonomics. Such variety of illumination directly affects the performance of computer vision algorithms, which tend to be sensitive to the visual appearance of faces and eyes in the images. In this work we present a robust DFW strategy that filters images and works seamlessly for direct IR. Extensive tests in a wide range of scenarios (day and night-time, presence of glasses, laboratory and field videos) show the promising results of this system.
BIB_text
title = {Robust Real-time Driver Drowsiness Detection System for Heterogeneous Lightning Conditions},
keywds = {
ADAS, computer vision, fatigue detection
}
abstract = {
Driver Fatigue Detection (DFW) has become the most active research area in the field of Advanced Driver Assistance Systems (ADAS). Its robust performance under real driving conditions requires the use of Infra-Red illumination. Direct IR illumination systems represent the current option for manufacturers and OEM, according to comfort and ergonomics. Such variety of illumination directly affects the performance of computer vision algorithms, which tend to be sensitive to the visual appearance of faces and eyes in the images. In this work we present a robust DFW strategy that filters images and works seamlessly for direct IR. Extensive tests in a wide range of scenarios (day and night-time, presence of glasses, laboratory and field videos) show the promising results of this system.
}
date = {2015-10-08},
year = {2015},
}
