Driver Fatigue Detection Method Based on Eye Multi-Feature Fusion

DOI

10.2991/iceea-18.2018.53How to use a DOI?

Keywords

face detection; random forest regression model; PERCLOS; CART; fatigue detection

Abstract

In order to improve the accuracy of driver fatigue detection, this paper proposes a new fatigue detection method based on random forest regression model and integrate eye fatigue surveillance method for multiple characteristic parameters. First, we use the features of simple class Haar to cascade algorithm Adaboost, aiming at carrying out quick face location detection. To solve the problem of low accuracy and high misclassification rate when uses the conventional method PERCLOS (percentage of eyelid closure over the pupil over time) via single parameter judgment, the paper increases four new parameters while extracting the eye fatigue characteristics and uses the classification and regression tree(CART) to judge whether the driver is awake or tired. The experimental results show that the accuracy rate of the driver's fatigue test is 96.7%, the error rate of the driver is 1.7% when the driver is in the fatigue state, the processing speed of our driver fatigue detection system is about 30 frames per second, it can make a judgment about the driver’s fatigue state in about 10 seconds, which owns higher accuracy and real-time.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Qinghua Liu
AU  - Haixiao Zhong
AU  - Xuehan Zhao
AU  - Lu Sun
PY  - 2018/03
DA  - 2018/03
TI  - Driver Fatigue Detection Method Based on Eye Multi-Feature Fusion
BT  - Proceedings of the 2018 2nd International Conference on Electrical Engineering and Automation (ICEEA 2018)
PB  - Atlantis Press
SP  - 241
EP  - 244
SN  - 2352-5401
UR  - https://doi.org/10.2991/iceea-18.2018.53
DO  - 10.2991/iceea-18.2018.53
ID  - Liu2018/03
ER  -