Vision-based Automated Fugl-Meyer Assessment System for Upper Limb Motor Function

The Fugl-Meyer Assessment (FMA) is an important method for evaluating upper limb motor functionality in stroke patients. However, traditional assessments by therapists not only increase the cost of labor and money, but also is subjective. In this paper, we propose a novel automated FMA system to overcome these limitations that patients cannot be accurately assessed in time. We use camera to capture the motion of the upper limb and then extract motion features. The MLP classification algorithm is developed to assign FMA scores using the features obtained from camera. Our work achieves average accuracy of 98.9% and 97.7% on a dataset of 15 volunteers and a dataset of S patients. Our result shows high accuracy and offers guidance to an automated assessment system.