Related Experiment Video
Updated: Jun 5, 2026
06:32
Bringing the Clinic Home: An At-Home Multi-Modal Data Collection Ecosystem to Support Adaptive Deep Brain Stimulation
Published on: July 14, 2023
Validating Single-Camera Pose Estimation Against Multi-Camera Motion Capture for Accessible Biomechanical Assessment
Aniket Pratapneni1,2, Ryan Halvorson1, Pavlos Silvestros1,3
1Department of Orthopaedic Surgery, University of California, San Francisco, San Francisco, CA 94122, USA.
IEEE Access : Practical Innovations, Open Solutions
|June 4, 2026
Summary
Single-camera pose estimation offers a low-cost, accessible alternative for motion analysis in musculoskeletal health. This study found it reliable for rehabilitation and telehealth, despite some accuracy limitations in dynamic movements.
Area of Science:
- Biomechanics
- Medical Technology
- Rehabilitation Engineering
Background:
- Musculoskeletal disorder diagnosis relies heavily on motion analysis.
- Traditional multi-camera systems are costly, environment-specific, and require expertise.
- Single-camera, deep-learning pose estimation presents a promising, accessible alternative.
Purpose of the Study:
- To evaluate the clinical accuracy and reliability of a single-camera pose estimation model (MeTRAbs).
- To compare MeTRAbs performance against a gold-standard multi-camera system (THEIA3D).
- To assess feasibility for rehabilitation and telehealth applications.
Main Methods:
- 51 participants performed gait, sit-to-stand, and trunk movements.
- Simultaneous recording using a smartphone camera (30Hz) and THEIA3D (180Hz).
- Root-mean-square errors (RMSE) and Intraclass Correlation Coefficients (ICCs) analyzed accuracy and reliability.
Main Results:
- Mean trajectory RMSE was 5.95 cm; joint angle RMSEs ranged from 2.10° to 10.98°.
- Proximal joints and frontal plane motions showed higher accuracy.
- Range of Motion ICCs exceeded 0.93, indicating strong reliability, with systematic bias identified as the main error source.
Conclusions:
- Single-camera pose estimation (MeTRAbs) demonstrates significant accuracy and reliability for clinical motion analysis.
- It is a feasible and scalable tool for accessible rehabilitation and telehealth.
- While not replacing multi-camera systems, it offers valuable insights for remote and outpatient settings.