Performance Evaluation of EMG MindRove Bracelet to Identify Hand Gestures to Control Self-adapted Serious Games

In neurological rehabilitation, EMG-controlled video games is an interesting alternative to complement classical treatments aiming to recover motor functionality. However, sometimes this type of applications are not ease to use due to the labour of EMG electrodes setup, being preferred sensors easy to wear. This article presents the performance analysis of the low-cost bracelet-type MindRove device in order to identify hand gestures to command a serious game. A hand gesture classifier is implemented in Python code. EMG signals were captured using a MindRove bracelet and an electrode-based system. The analysis show that the MindRove sensor includes noise that difficult the gesture identification and suggest that a larger sample is required.