A systematic review on surface electromyography-based classification system for identifying hand and finger movements

The developments in engineering fields have extended the use of electromyography (EMG) beyond traditional diagnostic applications to multifarious areas like movement analysis. Surface EMG-based gesture recognition systems can provide the instinctive and exact recognition of various gestures with an effective classifier. Many researches have been done intensively on recognizing wrist and whole-hand gestures. Still, very little research was done on individual finger gestures, as this is considered more challenging due to the complexity and subtleness of muscle for individual finger movements. In this study, we tried to review surface EMG-based hand/finger movement recognition techniques using machine learning classifiers. We selected and evaluated 33 primary documents for this systematic literature review according to the Kitchenham methodology. Based on the research questions, this paper analyzes machine learning algorithms’ applicability, accuracy, and efficiency. It reviews the basic hand/finger gesture recognition techniques using a standard model. We also tried to identify the trends and gaps in the studied articles that could lead to new areas of study in the future. •We adopted a systematic review to identify hand gestures using electromyography-based machine learning classifier.•We surveyed the hand gesture recognition process and proposed a standard recognition model.•Artificial neural networks show the highest accuracy for gesture recognition, and support vector machines were mostly used classifiers.•Overview the performance metrics for a suitable classifier that perform well with better accuracy.