EEG Feature Extraction of Motor Imagery Based on WT and STFT

摘要

Aiming at the problem of low recognition rate in classification task of motion imagery, an EEG feature extraction algorithm based on Wavelet Transform (WT) and Short-Time Fourier Transform (STFT) is proposed. Firstly, the wavelet decomposition of the EEG on a specific lead is used to reconstruct the frequency containing the Event-Related Desynchronization / Synchronization (ERD / ERS) features to remove the redundant information. Then the Short-Time Fourier Transform is used to extract the feature of the motor imagery, and visualized these features. Finally, the convolution neural network is used to classify and obtain the final result. The proposed method is applied to EEG data of two kinds of motion imaging tasks in BCI competition, and the experimental results show that the recognition rate of classification can reach 96.67% and the average Kappa coefficient is 0.93, which verifies that our proposed algorithm can effectively distinguish two types of motor imagery task, and improve the recognition rate of classification.