Hand Movement Recognition based on Empirical Mode Decomposition

R.D. Pinzón-Morales, Á.Á. Orozco-Gutierrez, K. Baquero-Duarte, and V.H.Grisales-Palacio (Colombia)


Empirical mode decomposition, electromyography, and pattern recognition


This paper presents a novel methodology for hand movement recognition based on empirical mode decomposition. Three-channel surface electromyographic (SEMG) signals are recorded using surface electrodes placed over the forearm for five different hand movements: closing, opening, supination, flexion and extension are considered in this document. A fixed-width window is used to segment each channel, then empirical mode decomposition is used to extract embedded oscillation in the EMG signal from which determinating features are extracted and then processed with different classifiers. The proposed method achieves a mean accuracy of 97.6% with a k-nearest neighbourhood classifier in the validation step.

Important Links:

Go Back