In this study, we introduce a novel variant and application of the Collaborative Representation based Classification in spectral domain for recognition of the hand gestures using the raw surface Electromyography signals. The intuitive use of spectral features are explained via circulant matrices. The proposed Spectral Collaborative Representation based Classification (SCRC) is able to recognize gestures with higher levels of accuracy for a fairly rich gesture set. The worst recognition result which is the best in the literature is obtained as 97.3\% among the four sets of the experiments for each hand gestures. The recognition results are reported with a substantial number of experiments and labeling computation.

Spectral Collaborative Representation based Classification for Hand Gestures recognition on Electromyography Signals