Real-Time Surface EMG Pattern Recognition for Hand Gestures Based on Support Vector Machine

In this paper, a real-time hand gesture recognition model based on surface electromyography signals is proposed. The Myo armband is applied to acquire surface electromyography (sEMG) signals generated by hand movement and we extract the features by using a sliding window to split the data. Then we select support vector machine (SVM) model for gesture recognition, which is trained by the training set and tested by testing set. We count the labels returned by classifier and assign the label whose times reach the threshold of recognition to the tested gesture. In order to evaluate this model, we set up an experiment with five target gestures which conducted by 12 subjects. Experimental results show that the average accuracy of proposed model can reach 97.8% and the recognition process is completed before the continuous action is finished, which means that this model can achieve real-time identification with high accuracy.