Real-Time static gesture recognition for upper extremity rehabilitation using the leap motion

Abstract

Cerebral Palsy is a motor disability that occurs in early childhood. Conventional therapy methods have proven useful for upper extremity rehabilitation, but can lead to non-compliance due to children getting bored with the repetition of exercises. Virtual reality and game-like simulations of conventional methods have proven to lead to higher rates of compliance, the patient being more engaged during exercising, and yield better performance during exercises. Most games are good at keeping players engaged, but does not focus on exercising finemotor control functions. In this paper, we present an analysis of classification techniques for static hand gestures. We also present a prototype of a game-like simulation of matching static hand gestures in order to increase motor control of the hand.