Corrective Suggestions to Exercise Performed at Home by Physiotherapy Patients using Motion Analysis.

Abstract

Physiotherapy patients pay a lot of money on their sessions where they do the exercises under an expert supervision. But not all can afford a physiotherapy appointment every now and then so few decide to redo the sessions at home by their own. These physiotherapy exercises are pose-sensitive, so in order to get maximum results, the workout pose should be as close to ideal as possible. Otherwise, all the effort put in the session would be for nothing and if done wrongly, it can cause injury also. So, there is the need of a benchmark system that can gauge the correctness of the workout pose performed by the patients remotely. This paper aims to address the issue and provide a system to solve it. The patient records the workout video and forwards it to the home workout benchmark system to get corrective suggestions. The system identifies the joint keypoints and the angles between all the involved keypoints. Further, with the application of time series data alignment algorithms, the system identifies the pose errors and generates the report to the patient. This system can be extended in future to the gyms and other health institutions for commercial use..