2019 · 20 citations · 12 references
ComputingComputer ScienceBiomedical Signal ProcessingFeature EngineeringSensor FusionElectromyogram AnalysisFeature ExtractionGesture AnalysisGesture RecognitionMuscle Signal ProcessingTime Series Feature ExtractionElectromyography-based Gesture RecognitionGesture Feature ExtractionHand Gesture AnalysisHand GesturesInformation RedundancyMomentsSensor-based Gesture RecognitionTime-domain Electromyography Features
This paper illustrates a collection of wide used time domain features of EMG signal, investigates their ability to distinguish hand gestures from one subject performing six different gestures with two sensor channels. The extracted set of features is presented as a fusion between the two sensor channels and a visual evaluation of their usability for performed hand gesture distinction is described. Results have shown that extracting Integrated EMG, Variance, Mean absolute values type one and type two, Average amplitude change and zero crossing time domain features are more significant for hand gesture recognition based on surface EMG signal analysis then extracting temporal moments coefficients, Wilson amplitude and myopulse percentage rate. However, the necessity of a feature selection step to eliminate information redundancy before hand gesture classification is concluded.
12