Abstract
So far, little is known how the sample assignment of surface electromyogram (sEMG) features in training set influences the recognition efficiency of hand gesture, and the aim of this study is to explore the impact of different sample arrangements in training set on the classification of hand gestures dominated with similar muscle activation patterns. Seven right-handed healthy subjects (24.2 ± 1.2 years) were recruited to perform similar grasping tasks (fist, spherical, and cylindrical grasping) and similar pinch tasks (finger, key, and tape pinch). Each task was sustained for 4 s and followed by a 5-s rest interval to avoid fatigue, and the procedure was repeated 60 times for every task. sEMG were recorded from six forearm hand muscles during grasping or pinch tasks, and 4-s sEMG from each channel was segmented for empirical mode decomposition analysis trial by trial. The muscle activity was quantified with zero crossing (ZC) and Wilson amplitude (WAMP) of the first four resulting intrinsic mode function. Thereafter, a sEMG feature vector was constructed with the ZC and WAMP of each channel sEMG, and a classifier combined with support vector machine and genetic algorithm was used for hand gesture recognition. The sample number for each hand gesture was designed to be rearranged according to different sample proportion in training set, and corresponding recognition rate was calculated to evaluate the effect of sample assignment change on gesture classification. Either for similar grasping or pinch tasks, the sample assignment change in training set affected the overall recognition rate of candidate hand gesture. Compare to conventional results with uniformly assigned training samples, the recognition rate of similar pinch gestures was significantly improved when the sample of finger-, key-, and tape-pinch gesture were assigned as 60, 20, and 20%, respectively. Similarly, the recognition rate of similar grasping gestures also rose when the sample proportion of fist, spherical, and cylindrical grasping was 40, 30, and 30%, respectively. Our results suggested that the recognition rate of hand gestures can be regulated by change sample arrangement in training set, which can be potentially used to improve fine-gesture recognition for myoelectric robotic hand exoskeleton control.
Highlights
Myoelectric control systems have been widely used to control assistive and rehabilitation devices, i.e., EMG-controlled robotic hand exoskeleton (Leonardis et al, 2015), which collected the surface electromyogram from the forearm muscles of non-paretic hand controlling the movement of exoskeleton, and to train and/or guide the grasping or pinch task conduction of paretic hand as well
Feature classification of surface electromyogram (sEMG) in time and/or frequency domain is usually employed for recognizing non-paretic hand gesture under the following principle: different hand motions/gestures are dominated with different muscle activity patterns, which result in a distinguishable sEMG feature vector (Lima et al, 2016)
This study is to investigate how sample arrangement in training set affects the hand gesture classification accuracy. sEMG signals have been recorded from forearm hand muscles when conducting similar grasping gestures or similar pinch gestures, and the impact of the sample proportion in the training set on the recognition efficiency of similar hand gestures are evaluated by changing the sample number of each candidate gesture
Summary
Myoelectric control systems have been widely used to control assistive and rehabilitation devices, i.e., EMG-controlled robotic hand exoskeleton (Leonardis et al, 2015), which collected the surface electromyogram (sEMG) from the forearm muscles of non-paretic hand controlling the movement of exoskeleton, and to train and/or guide the grasping or pinch task conduction of paretic hand as well. A variety of myoelectric pattern identification strategies have been proposed to classify the sEMG signals for different hand gestures, very little attention has been paid to the recognition of hand gestures dominated with similar hand muscle activity patterns (AbdelMaseeh et al, 2016). As one of the most dexterous organs in the world, our hand can perform a variety of hand motions with different finger coordination patterns, and part of these hand motions are controlled with almost same hand muscle contraction patterns, such as hand pinch and hand tripod gestures These hand gestures with similar muscle activities patterns were usually excluded from hand motion classification studies due to their low sensitivity and poor classification performance (Castro et al, 2015). To train the paretic hand after stroke with a robotic hand exoskeleton, it is necessary to identify gestures with high similarity based on sEMG features detection from contralateral non-paretic hand
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
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
