Data fusion based portable lightweight human brain state detection method

Abstract

The invention discloses a data fusion based portable lightweight human brain state detection method. The method includes: acquiring original electroencephalogram signal data of N channels by electroencephalogram signal acquisition equipment, and preprocessing the original electroencephalogram signals data; subjecting the preprocessed original electroencephalogram signals data to blind source separation to obtain signals of multiple signal sources, and performing feature extraction on the signals of each signal source on the basis of wavelet packet transform; inputting each signal source into aplurality of well trained lightweight convolutional neural network models to analyze, and subjecting outputs of the multiple lightweight convolutional neural network models to weighted voting to obtain a final classification result. The lightweight convolutional neural network models take features obtained through wavelet packet transform of each signal source as inputs and take signal source types as outputs.

Description

technical field [0001] The invention belongs to the technical field of brain-computer interface, and in particular relates to a portable lightweight human brain state detection method based on data fusion. Background technique [0002] At present, people's demand for physiological state monitoring is increasing day by day. It is very important to use EEG signals to monitor people's physiological state. In the traditional monitoring based on EEG signals, the characteristics of the signal are first extracted by time-frequency analysis method, and then the signal is analyzed by machine learning methods such as SVM and k-means. However, the final analysis accuracy of these methods is not ideal. With the emergence of deep learning, CNN, RNN and other methods have also performed well in the analysis of EEG signals. However, due to the high structural risk of the deep learning model, the model is prone to poor generalization ability, over-fitting, and poor real-time performance....

AI Technical Summary

Problems solved by technology

[0003] The technical problems to be solved by the present invention are: the low accuracy of EEG signal analysis based on a small number of channels and the poor training time and real-time response time of the analysis model based on deep learning, as well as the fact that the analysis of EEG signals in the past used many The problem of inconvenient portability

Images