CSP Motor Imagery Classification Apparatus and Method thereof using Cross-channel-region CSP features

摘要

Disclosed is a motor imagery classification method. The motor imagery classification method comprises: (a) a step performed by an area classification unit included in a motor imagery classification apparatus to divide a plurality of electrodes attached to a subject into k areas each including n electrodes; (b) a step performed by a partial electroencephalogram signal generation unit included in the motor imagery classification apparatus to couple an electroencephalogram signal of a motion to be classified measured from a target electrode which is any one electrode among the plurality of electrodes to an area electroencephalogram signal of the motion to be classified measured in any one area among the k areas to generate k partial electroencephalogram signals per target electrode; (c) a step performed by a feature vector extraction unit included in the motor imagery classification apparatus to apply a common spatial pattern (CSP) technique to the partial electroencephalogram signals to extract a feature vector; and (d) a step performed by a classification unit included in the motor imagery classification apparatus to use the feature vector to learn a classifier.