Convolutional neural network based features for motor imagery EEG signals classification in brain–computer interface system

摘要

One of the essential challenges in brain–computer interface is to classify motor imagery (MI) signals. In this paper, anensemble SVM-based voting system is proposed. In each line of this system, the EEG signal is transformed into differentrepresentations based on discrete cosine transform, Fourier transform, common spatial pattern, and empiricalmode decomposition, and then these representations are combined in a triple-frame matrix. These frames are fed intoa pre-trained deep convolutional neural network as a feature extractor. For each line, an SVM is employed to classify theextracted feature vectors. Finally, a decision is made based on voting between these SVMs. Performance of the proposedmethod is examined on the BCI Competition III dataset Iva to separate right hand and foot movement imagery. The simpleproposed method achieves the average accuracy of 96.34% for all of the subjects, and 99.70% for the best situation thatis an improvement in MI classification. In addition, it can be seen that right side of the brain is more effective than theother side in EEG-based MI classification.