Presenter of 1 Presentation
AUTOMATION OF FUGL-MEYR ASSESSMENT
Abstract
Background and Aims
Majority of Stroke patients are afflicted with Hemiparesis, characterized by weakness and partial loss of function in one side of the body. To monitor motor recovery post Stroke, the Fugl-Meyer assessment (FMA) is the most popular index for evaluating upper extremity motor function in stroke patients. In order to overcome inter-physician variability in scores and establish more frequent assessments at low-cost, we propose to automate FMA using the power of Artificial Intelligence.
Methods
The system comprises of a software which can be deployed on a medium-end computer that captures a video of the patient performing the prompted movement using a webcam, detects the joints with the help of Google’s Mediapipe pose estimation model, processes the localized joints frame by frame using a rule-based classifier, and returns the score for the corresponding movement.
Results
For efficient extraction of features, the subject is to be seated 2m away from the camera and at an orientation of either 0 or 45 degrees depending on the class of movement. The proposed system has been able to process the sequentially detected joints and extract the desired features for scoring of 24 movements of Upper-extremity.
Conclusions
The automated, cost-efficient framework aids the patient to monitor their progress more frequently from their homes. In addition to the qualitative FMA scores, the extracted features such as movement time, velocity, range of motion provides a quantitative estimate of improvement over time. Thus, the Physician can alter the course of rehabilitation by monitoring the performance and needs of the patient flexibly and remotely.