Pattern Recognition of Motion Imagination EEG Signal Based on Convolutional Cyclic Neural Network
Brain-computer interface technology helps people with motor disorders interact with the environment through external devices.In order to improve the pattern recognition rate of EEG signals fuelled by motion imagination,a hybrid neural network pattern recogni-tion based on convolutional neural network(CNN)and recurrent neural network(RNN)is proposed,In the actual calculation,two different RNNs,Long Short-Term Memory(LSTM)and Gated Recurrent Unit(GRU),are used for experimental comparison.Firstly,the original EEG data is filtered and segmented and the processed data are input into the hybrid neural network;Finally,Softmax is used for classification.Two EEG data sets,data set 2a and data set 1 in BCI competition Ⅳ,are used for experimental verification.The proposed methods can effectively improve the accuracy of pattern recognition,with an average accuracy more than 95%.
万方数据
维普
