Evaluation on EMG Electrode Reduction in Recognizing the Pattern of Hand Gesture by Using SVM Method

• Published in: Journal of physics. Conference series Vol. 1577; no. 1; pp. 12044 - 12050
• Main Authors: Winarno, H A, Poernama, A I, Soesanti, I, Nugroho, H A
• Format: Journal Article
• Language: English
• Published: Bristol IOP Publishing 01.07.2020
• Subjects: Electrodes; Electromyography; Feature extraction; Feature recognition; Gesture recognition; Pattern recognition; Physics; Prostheses; Reduction; Support vector machines; Wrist
• ISSN: 1742-6588; 1742-6596
• DOI: 10.1088/1742-6596/1577/1/012044

Understanding the pattern of hand gesture on research which designs a prosthetic hand has been popular subject in recent years. The hand gesture recognition relies heavily on the sensors, with electromyograph (EMG) sensor electrode being used the most. Previous researches have been using various recognition methods with different number of electrode. But, many of them used 12 to 4 electrodes. In this paper, we experimented the use of 3 to 2 electrodes to recognize hand gesture in open (HO), close (HC) wrist flexion (WF), and wrist extension (WE) condition. It used SVM method to recognize the extracted features from electrode signal. In this research, two scenarios between the use of 3 electrodes and then with the reduction to 2 electrodes are compared. SVM classification showed 96.35% as the best accuracy on experiment that used 3 electrodes and 97.16 % when it used 2 electrodes. Mean, standard deviation, and root mean square were being the best statistical features in recognizing HO, HC, WE and WF hand gesture.