Wavelet transform and singular value decomposition of EEG signal for pattern recognition of complicated hand activities

Abstract

Electroencephalogram (EEG) is a useful bioelectrical signal in pattern recognition of hand activities because of its following characteristics: (1) patterns of EEG are different when doing diverse movements and mental tasks; (2) EEG can be real-timely or instantly extracted; (3) the measurement of EEG can reach an enough precision. A new approach for the pattern recognition of four complicated hand activities based on EEG is presented in this paper, in which each piece of raw data sequence for EEG signal is decomposed by wavelet transform (WT) to form a matrix, and the singular value of the matrix is extracted by singular value decomposition (SVD). And then the singular value, as the feature vector, is input to the artificial neural network (ANN) to discriminate the four hand activities including grasping a football, a small bar, a cylinder and a hard paper. Finally the research results show the correct classification rate of 89% was achieved by the approach mentioned above. © Springer-Verlag Berlin Heidelberg 2007.