The development of a Cluster-based Motion Capture system for Upper-limb Stroke Rehabilitation.

This paper presents the development of a cluster based, motion capture enabled, rehabilitation game for intensive practice in stroke rehabilitation. The implementation uses a low cost 6 camera Vicon Origin system and their Evoke software. The data is streamed live to a Visualstudio2022 bespoke c++ program (SCM) which implements a biomechanical cluster based model of the upper arms utilizing a pointer and 12 key anatomical landmarks. The data is then lived streamed to Unreal Engine 5 where it is used as the input to a 3D bubble popping game designed for stroke rehabilitation in mild to moderately effected stroke survivors. The system was successfully implemented and gave reliable data to drive the gamification. The system is expensive and is easy to setup in a clinic, leisure centre or other community rehabilitation space. The game can be progressed in difficulty by increasing the required volume of action, by increasing the speed of appearance of targets and by increasing the speed of movement of the targets. Further development of the system will see it tested with stroke survivors as part of an 8-week rehabilitation program beginning in September 2023 and its extension to other more complex and cognitively challenging tasks and games. The software used (SCM and Unreal Engine 5 program) will be made publically available free of charge for details please contact the lead author.