Motion Pattern Tracking for Home Based Stroke Rehabilitation Exercise Using MESA SR4500

An automated physiotherapy exercise monitoring system requires good postural structure and pose information. Most of human movements for daily activity pivoted on lower limb that requires intensive care, especially for the stroke patient. However, there are a lot of uncertainties in human stance, movement and strength that need to be observed for good rehabilitation training. Hence, home-based rehabilitation monitoring system for lower limb therapy is crucial to encourage and facilitate the patient to perform the exercise effectively. This paper presents a novel real time rehabilitation system by tracking the lower limb information based on 3D information by using MESA SR4500, which produces Red, Green, Blue and Depth (RGBD) information. In this paper, only Depth and grayscale image have been utilized to track and evaluate the patient movement. A marker that located at the centroid of the moving body region has been estimated by using Kalman Filter, which acts as the main component. The pattern for a correctly performed exercise is different compared to the wrongly done exercise. The motion pattern has been evaluated to test the effectiveness of the exercise from the obtained graph. In conclusion, the results show that depth information is more important compared to grayscale information as it affects the tracking performance the most.