Related Experiment Video
Updated: May 20, 2025
04:37
Comprehensive Understanding of Inactivity-Induced Gait Alteration in Rodents
Published on: July 6, 2022
2.3K
Reliability and validity of current computer vision based motion capture systems in gait analysis: A systematic
Xingye Cheng1, Yiran Jiao1, Rebecca M Meiring2
1Department of Exercise Sciences, Faculty of Science, University of Auckland, Auckland 1023, New Zealand.
Gait & Posture
|April 18, 2025
Summary
Pose estimation algorithm (PEA)-based gait analysis offers an accessible alternative to traditional methods. This review evaluates PEA accuracy, finding it depends on factors like camera settings and algorithms, with varying results for different joint angles.
Area of Science:
- Biomechanics
- Computer Vision
- Rehabilitation Technology
Background:
- Traditional instrumented gait analysis (IGA) is costly and complex.
- Pose estimation algorithms (PEAs) offer an accessible video-based approach for gait analysis.
- Limited review exists on PEA reliability, validity, and accuracy factors in gait analysis.
Purpose of the Study:
- To systematically review the accuracy of PEA-based gait analysis systems.
- To identify algorithmic factors influencing PEA accuracy in gait analysis.
Main Methods:
- Systematic literature search across Scopus, PubMed, and IEEE.
- Inclusion and exclusion criteria applied to 644 initial articles, resulting in 20 reviewed.
- Extraction of reliability, validity, and algorithmic parameters for analysis.
Main Results:
- Most studies focused on validity; reliability evidence is limited.
- OpenCap showed good accuracy for 3D joint angles (MAE 4.1°), but rotational angles need more validation.
- OpenPose demonstrated high accuracy for spatiotemporal and sagittal plane hip/knee angles, but poor accuracy for ankle kinematics.
- PEA accuracy is influenced by camera settings, backbone architecture, and training data.
Conclusions:
- PEA-based gait analysis shows promise but requires further validation, especially for ankle and rotational joint kinematics.
- Algorithmic and technical factors significantly impact PEA accuracy.
- This review supports further research into PEA for clinical gait analysis and rehabilitation.
