Fatigue Classification Detection of Single-Channel EEG Signal
To solve the problem of weak scalp EEG potential signals,susceptibility to interference,non stationarity and randomness,and difficulty in manually extracting features,we propose a CNN+BiGRU network model that fully extracts the correlation information between EEG signal sequences before and af-ter.The experimental results show that our proposed dual stream network model achieves an accuracy of 95.70%.Compared with existing research methods,the accuracy and feasibility of fatigue detection using single channel EEG signals have been significantly improved,providing new ideas for fatigue detection re-search.
万方数据
维普
