Hand Gesture Recognition Based on Surface EMG Using Feature Fusion and Machine Learning Approaches

Abstract-Surface electromyography hand gesture recognition has received a lot of interest, particularly in the domain of human-computer interfaces and intelligent natural rehabilitation in the last several years. Previous research has provided numerous algorithms for recognising gesture classes using sEMG signals, however, their recognition performance is affected by the following factors: (i) the extracted features are not distinguishable to recognize closely related gestures, (ii) with an increase in feature dimension, the recognition model gets confused to classify the correct class, (iii) the performance degrades for higher number of gesture classes. In this paper, several distinguished hand-crafted time domain features such as mean absolute value (MAV), standard deviation (SD), and waveform length (WL) features, are derived from sEMG signals to represent a gesture class. The feature combinations are evaluated using a variety of machine learning classification models. The proposed features and RF classifier outperform prior reported findings on Ninapro- DB5 standard dataset, with an accuracy of 79.96%.