Transfer Components Between Subjects for EEG-based Driving Fatigue Detection

In this paper, we first build up an electroencephalogram (EEG)-based driving fatigue detection system, and then propose a subject transfer framework for this system via component analysis. We apply a subspace projecting approach called transfer component analysis (TCA) for subject transfer. The main idea is to learn a set of transfer components underlying source domain (source subjects) and target domain (target subjects). When projected to this subspace, the difference of feature distributions of both domains can be reduced. Meanwhile, the discriminative information can be preserved. From the experiments, we show that the TCA-based algorithm can achieve a significant improvement on performance with the best mean accuracy of 77.56 %, in comparison of the baseline accuracy of 66.56 %. The improvement shows the feasibility and efficiency of our approach for subject transfer driving fatigue detection from EEG.