Hand Movement Recognition Based on Singular Value Decomposition of Surface EMG Signal

Surface electromyography (sEMG) signals represent electrical activity of muscle cells and are extensively used for prosthetics development. In this paper, an effective technique is put forward to classify some typical hand movements from the sEMG signals based on singular value decomposition (SVD) and principal component analysis (PCA). In view of employing the SVD on a frame of sEMG data, first, short duration overlapping sub-frames are extracted to form a sub-frame matrix. We propose to employ the SVD on the sub-frame matrix to extract singular values as well as principal components avoiding computation involved in the PCA. Apart from the extracted singular values, some statistical parameters of the first five principal components are proposed as features seeing as, in the eigenspace, the projected values of the original data are expected to offer more distinguishable characteristics for different hand movements. With a view to performing the classification, the K-nearest neighborhood (KNN) classifier is applied in a hierarchical approach. The suggested technique is put to the test considering 5 cross 2 cross validation scheme on a publicly available sEMG database consisting of six different hand movements obtained from three females and two males. It is found that the proposed technique offers consistently high classification accuracy in classifying various hand movements with lower computational complexity.