Multi-Class Motor Imagery Classification

Motor imagery (mi) classification task is a high dimension multivariate and complicated subject. in this respect, the original signals are analyzed and minimal unique features of the classes are extracted to facilitate accurate classification of the actions performed. the fusion of common spatial pattern, fisher discriminate ratio and filterbank alongside with the svm and cnn-lstm classiers are incorporated to provide accurate grouping. as a result and after extensive simulations, it is shown that the csp+ fdr + cnn-lstm setup more accurately differentiates among classes.