Vague Gesture Control: Implications for Burns Patients

In the rehabilitation of burns patients, remedial exercises are an important part of maintaining and regaining range-of-motion. To encourage patient rehabilitation and reduce discomfort, this paper presents the preliminary stages of a project which makes use of both vision based sensors and virtual reality technologies to create an immersive environment. The goal of our research is to design and develop effective vague gesture control. Hand gesture detection may lack clarity/resolution or otherwise be vague due to various reasons including patients' difficulty to enact gestures fully due to injuries, interference with the visual detection of gestures due to bandages, and the difficulty that vision based sensors have detecting small details like fingers at range. Extra environmental and situational information can enhance gesture recognition and provide better gestural classification. To achieve this, temporal machine learning is used to develop an application which uses past environmental interactions to help clarify present vague gestures. A gesture recognition system that can handle imperfect gestures in an intuitive way contributes towards naturalistic computer human interactions, in particular for gamified burns rehabilitation systems.