Interpretable Driver Fatigue Estimation Based on Hierarchical Symptom Representations

Jiaqin Lin,Shaoyi Du,Yuying Liu,Zhiqiang Tian,Ting Qu,Nanning Zheng

DOI: https://doi.org/10.1007/978-3-031-27818-1_53

2023-01-01

Abstract:Traffic accidents caused by driver fatigue lead to millions of death and financial loss every year. Current end-to-end methods for driver fatigue detection are not capable of distinguishing the detailed fatigue symptoms and interpretably inferring the fatigue state. In this paper, we propose an interpretable driver fatigue detection method with hierarchical fatigue symptom representations. In pursuit of a more general and interpretable driver fatigue detection approach, we propose to detect detailed fatigue symptoms before driver state inferring. First of all, we propose a hierarchical method that accurately classifies abnormal behaviors into detailed fatigue symptoms. Moreover, to fuse the fatigue symptom detection results accurately and efficiently, we propose an effective and interpretable fatigue estimation method with maximum a posteriori and experience constraints. Finally, we evaluate the proposed method on the driver fatigue detection benchmark dataset, and the experimental results endorse the feasibility and effectiveness of the proposed method.