| Popis: |
Aiming to the ERD/ERS phenomenon of left-right hand imaginary movement, this paper presents a method of wavelet transform combined with statistical analysis to extract EEG features evoked by imaginary movement. And the features were classified using the support vector machine based on RBF kernel and cross-validation accuracy (CVA) method. The results have shown that this method can perform effectively to extract features and reflect ERS and ERD characteristics of EEG signal. The accuracy of classification can reach 90% within the time costing 3.5 seconds. The highest signal to noise ratio is 1.445, and the maximum mutual information is 0.645bit. The results can meet the real-time brain-computer interface system. |
