Mental Tasks Signal Classification

Electroencephalogram (EEG) signal is animportant source of information for knowingbrain processes. To interpret the brain activity,Matching Pursuit Based EEG signalclassification is proposed. This system includesthree main components which are Preprocessing,Feature extraction and Classification. In thepreprocessing step, Wavelet Packet IndependentComponent Analysis (WPICA) method is used toremove some unwanted noise of EEG recording.Matching Pursuit (MP) with Wavelet PacketDictionary is used to extract the features of EEGsignal. The k Nearest Neighbor (kNN) classifiedthe extracted MP features. In this work, theKeirn and Aunon EEG dataset is used in theexperiments. The feature extracted from MPbased wavelet packet dictionary achieved over90% accuracy in two seconds length ofbrainwave signal in five mental tasksclassification