Enhancing Physical Therapy through Motion Tracking: A Kinect-based Approach

Zhou, Kelly [Browse]
Senior thesis


Kaplan, Alan [Browse]
Princeton University. Department of Computer Science [Browse]
Class year
Summary note
People of all ages rely on physical therapy to improve their physical health. Physical therapy helps people recover from injuries, cope with chronic musculoskeletal diseases, and increase strength and flexibility in old age. Of the millions of physical therapy patients in the United States each year, however, only a small percentage are fully compliant with their plan of care. This low adherence stems largely from unintentional malpractices of their exercises as many patients are unable to follow their prescribed exercises accurately outside of the therapist's office. Since hands-on guidance from trained physical therapists proves difficult to scale, we propose a technological solution. For this project, we apply consumer-based technology, specifically the Microsoft Kinect, to track motion and assess accuracy of physical therapy exercises practiced at home. We propose a solution that tracks a user's movements, assesses accuracy, and provides instantaneous feedback. For the scope of this project, we focus primarily on physical therapy practices catered toward older demographics, namely exercises designed to increase balance and strength. Following an evaluation with subject matter experts, we found that the system accurately assessed key form and body alignment for three major exercises—the squat, the single leg stance, and the heel raise—and thus proves promising for improving patient compliance with at-home exercise programs. The system can be enhanced for future use by improving visual, and adding audio, feedback for users and increasing overall precision of measurements.

Supplementary Information