Toward Accessible Remote Physiotherapy for Older Adults: A Feasibility Study of a Markerless Motion Tracking System Using Adapted Consumer-Grade Sensors

Read full text Summarise with VIRAT
Authors: Jeff Gajdacs, Qiyin Fang, Pasqualina L. Santaguida, Mhan Pouromidi.
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Purpose Markerless motion tracking offers a promising approach to remote physiotherapy, particularly for older adults and individuals in low-resource settings.However, existing systems are often optimized for gait analysis and lack validation for therapeutic exercises [1].This study aimed to adapt a consumer-grade, privacy-focused sensor (the AltumView Sentinare 3) into a low-cost, dual-camera system for geriatric rehabilitation monitoring, and to conduct a preliminary feasibility evaluation.Method The system was configured in a controlled laboratory setting using two orthogonally positioned AltumView cameras.Preliminary kinematic data were collected from three researchers (ages 24-35) performing standardized geriatric-relevant exercises (simple step-ups, partial squats).A custom Python pipeline retrieved anonymized 2D joint coordinates via a secure API, using a framework conceptually similar to methods like OpenPose [2], reconstructed 3D kinematics through vector triangulation, and calculated joint angles (knee, hip, spine), angular velocities, and accelerations.Validation included synchronized qualitative video recording and systematic logging of tracking errors, data dropouts, and measurement consistency.Performance was evaluated in terms of occlusion resilience, setup time, and practical deployment feasibility.Results and Discussion Initial testing demonstrated the system's ability to generate continuous 3D joint angle profiles during multiplanar movements relevant to geriatric rehabilitation.Tracking was robust under clear viewing conditions, with skeletal data successfully captured for >95% of frames during exercises such as step-ups.Data loss occurred primarily during severe self-occlusion (e.g., cross-body movements), highlighting a known limitation of monocular vision-based systems.The system's sampling rate (~5 fps) limited temporal resolution but remained sufficient for slow therapeutic exercises.Practical advantages included rapid setup (

Fields of Study & Subjects

Human Pose and Action Recognition Context-Aware Activity Recognition Systems Stroke Rehabilitation and Recovery

Publication details

Journal / source
Gerontechnology
Pages
1–1
Publication date
1 January 2026
Language
EN