Related Experiment Video
Updated: Jul 23, 2025
07:51
Video Movement Analysis Using Smartphones ViMAS: A Pilot Study
Published on: March 14, 2017
16.9K
Automated Implementation of the Edinburgh Visual Gait Score (EVGS) Using OpenPose and Handheld Smartphone Video.
Shri Harini Ramesh1, Edward D Lemaire2,3, Albert Tu4
1Department of Mechanical Engineering, University of Ottawa, Ottawa, ON K1N 6N5, Canada.
Sensors (Basel, Switzerland)
|July 11, 2023
Summary
An AI algorithm automatically scores human movement using smartphone videos, making gait analysis more accessible. This automated Edinburgh Visual Gait Score (EVGS) method offers a faster, cost-effective alternative to manual video analysis.
Area of Science:
- Biomechanics and Movement Science
- Computer Vision and Artificial Intelligence
- Digital Health
Background:
- Quantitative human movement evaluation via digital video is advancing with AI.
- The Edinburgh Visual Gait Score (EVGS) is valuable but requires time and expertise for manual scoring.
- Current methods for gait analysis can be time-consuming and resource-intensive.
Purpose of the Study:
- To develop an automated algorithmic implementation of the EVGS using smartphone video.
- To enable accessible and cost-effective gait analysis through AI.
- To reduce the time and expertise needed for gait assessment.
Main Methods:
- Video recording of participants' walking using a smartphone at 60 Hz.
- Identification of body keypoints using the OpenPose BODY25 pose estimation model.
- Development of an algorithm for foot event and stride detection, and EVGS parameter calculation.
Main Results:
- Accurate stride detection within 2-5 frames.
- Strong agreement between algorithmic and human EVGS scores for 14 of 17 parameters.
- High correlation (r > 0.80) of algorithmic EVGS results with ground truth for 8 of 17 parameters.
Conclusions:
- An AI-driven approach using smartphone video can automate EVGS scoring effectively.
- This technology enhances the accessibility and cost-effectiveness of gait analysis.
- Findings support future research in remote gait analysis using AI and mobile devices.
