A DATASET AT A GLANCE...

Dataset for EMG-based Gesture Recognition… 

Electromyography (EMG), similar to Electroencephalogram (EEG), is a biomedical signal that signifies the muscle responses on nerve stimulation. In addition to clinical diagnostics of several neuromuscular disorders, EMG does possess wider applications. Few of them include: (i) Human- Computer Interaction (HCI), (ii) Prosthetics and Orthotics, (iii) Gait Analysis, (iv) Examining occupational health of the employees and (v) Biomechanics as well as sports science. Since all these applications have a great impact on the current technological era, various researchers have been working on varied EMG- based applications. However, when speaking about gesture recognition using EMG in particular, the researchers can make use of the ‘EMG data for gestures’ from the UCI Machine Learning Repository (https://archive.ics.uci.edu/dataset/481/emg+data+for+gestures). This dataset owns about 40000-50000 recordings, of which 30000 are completely assured for application. The creators of the dataset have worn a MYO Thalmic bracelet on a user’s forearm, which has equally-spaced eight sensors over it for capturing the myographic signals. Further, these signals have been recorded to a PC through a Bluetooth receiver.   A researcher can use the eight sensor recordings at ‘t’ times in milliseconds to develop a gesture recognition framework, which classify the EMG signals to be one among the following eight classifications: unmarked data, hand at rest, hand clenched in a fist, wrist flexion, wrist extension, radial deviations, ulnar deviations and extended palm (implying that the entire number of subjects have not performed any gesture).  An user of this dataset might also cite the dataset as “Krilova,N., Kastalskiy,I., Kazantsev,V., Makarov,V.A., and Lobov,S.. (2019). EMG data for gestures. UCI Machine Learning Repository. https://doi.org/10.24432/C5ZP5C”. Any researcher with the vision to use gesture recognition in medical and human computer interactions is best-supported with this dataset.