Spatiotemporal gait analysis using video from a moving observer.

Area of Science:

• Biomechanics
• Human Movement Analysis
• Wearable Technology

Background:

• Spatiotemporal gait analysis is crucial for assessing human movement in clinical and research settings.
• Existing methods often lack simplicity and versatility for real-world applications.
• A need exists for accessible and scalable gait analysis solutions.

Purpose of the Study:

• To develop and evaluate a novel method for spatiotemporal gait analysis using smart glasses worn by an observer.
• To assess the accuracy and reliability of this observer-based method in diverse real-world conditions.

Main Methods:

• The study employed monocular 3D pose estimation and visual-inertial odometry.
• Data were collected from 608 walking sequences across varied indoor and outdoor environments.
• A reference system combining inertial and video-based motion capture was used for validation.

Main Results:

• The proposed method achieved high accuracy in detecting key gait events (heel-strike, toe-off) with a mean error of 8.7 ± 37.3 ms.
• Moderate to excellent agreement was observed for both temporal (ICC: 0.74–0.98) and spatial (ICC: 0.87–0.99) gait parameters.
• Performance remained consistent across various environments, slopes, trajectories, viewpoints, and viewing distances.

Conclusions:

• Observer-based gait analysis using smart glasses is a viable and accurate method for obtaining spatiotemporal parameters.
• This approach offers a simple, scalable, and reliable alternative to traditional gait analysis techniques.
• The findings support the use of this technology for real-world human movement assessment.