Search for: All records

Current kinematic analysis for patients with upper or lowerextremity challenges is usually performed indoors at the clin-ics, which may not always be accessible for all patients. Onthe other hand, mobility scooter is a popular assistive toolused by people with mobility disabilities. In this study, weintroduce a remote kinematic analysis system for mobilityscooter riders to use in their local communities. In order totrain the human pose estimation model for the kinematic anal-ysis application, we have collected our own mobility scooterriding video dataset which captures riders’ upper-body move-ments. The ground truth data is labeled by the collaboratingclinicians. The evaluation results show high system accuracyboth in the keypoints prediction and in the downstream kine-matic analysis, compared with the general-purpose pose mod-els. Our efficiency test results on NVIDIA Jetson Orin Nanoalso validate the feasibility of running the system in real-timeon edge devices.