Cross-subject eeg fatigue state classification method based on generative adversarial domain adaptation

Abstract

The invention discloses a cross-subject EEG fatigue state classification method based on generative confrontation domain self-adaptation. The invention first acquires data and preprocesses to remove artifacts; secondly, extracts EEG features through PSD, and obtains a two-dimensional sample matrix from three-dimensional EEG time series; then distinguishes source domain and target domain data sets to obtain non-overlapping training sets and The test set; then use part of the unlabeled target domain data and random data that conforms to the Gaussian distribution to train the classification model GDANN; finally use the confusion matrix to evaluate the accuracy of the classification results. The invention further combines the idea of ​​generative confrontation network and domain invariance, which not only solves the problem that the EEG signal data set is scarce and difficult to obtain, but also balances the problem of mismatching source domain data and target domain data, and avoids negative transfer to a certain extent. A high-precision cross-subject fatigue detection classifier has been trained, with a view to having a broad application prospect in actual brain-computer interaction.

Description

technical field [0001] The invention belongs to the field of electroencephalogram signal (EEG) fatigue state recognition in the field of biological feature recognition, and in particular relates to a cross-subject EEG fatigue state classification method based on generational confrontation domain self-adaptation. Background technique [0002] Mental fatigue is a complex physical and psychological state that can lead to decreased alertness, concentration and cognitive abilities. About 1.3 million people die in traffic accidents every year in the world, and fatigue driving is the main factor therein. Therefore, how to effectively detect and predict the mental state while driving is very important for reducing the loss of life and property caused by fatigue driving. [0003] In recent years, many fatigue driving detection methods have been proposed. Among them, the mental state analysis method based on EEG is an effective and objective fatigue detection method. Because the EE...

AI Technical Summary

Problems solved by technology

[0006] In order to overcome the problems of poor discriminative performance in cross-subject EEG fatigue state classification of the above-mentioned prior art, data mismatch in source domain and target domain, and negative transfer caused by too many source domain samples, the present invention proposes a method based on generative confrontational domain adaptive Cross-subject EEG fatigue state classification method Generative-DANN (GDANN)

Images