Abstract
A hand gesture recognition using surface electromyography signals can be applied in several scenarios such as prosthetic limb control, human computer interface and so on. In this paper, we propose a hand gesture recognition method using single-channel electrodes to classify 5 hand gestures: Pinch, Fingers spread, Fist, Wave in, and Wave out. The proposed method is consisted of 4 steps: preprocessing, data segmentation, feature extraction and classification. At preprocessing step, a muscle activity detection algorithm is applied to remove rest parts of the dataset. At data segmentation step, an overlapped windowing algorithm is applied to segment the main window for acquiring sub-windows. At feature extraction step, one part of features is observed from above mentioned sub-windows. And time domain features of the raw surface electromyography signals are extracted as another part. Finally for classification, an artificial neural network is modelled to label sub-window observation. Experiment shows the proposed method has an accuracy of 91.3%.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call