Related Experiment Video
Updated: Jul 12, 2026
Video Movement Analysis Using Smartphones (ViMAS): A Pilot Study
Published on: March 14, 2017
MobileGait: A low-cost camera-based mobile application for gait assessment in elderly adults
Yutong Du1, Arthur Tay1, Jiun Shiah Low1
1Department of Electrical and Computer Engineering, College of Design and Engineering, National University of Singapore, Singapore.
Background:
Gait assessment provides key indicators of mobility, neurological function, and fall risk. It is widely used to monitor rehabilitation and disease progression. However, conventional motion-capture systems are costly, take up space, and require expert operation. Recent advances in pose estimation have enabled camera-based gait analysis, yet most implementations remain research prototypes that rely on manual intervention or offline processing. Fully automated and deployable assessment pipelines integrating data capture, processing, and reporting remain limited.
Objective:
This study presents and validates the accuracy of MobileGait, a low-cost, camera-based mobile application for automated gait assessment that integrates mobile data acquisition, server-side processing, and web-based reporting.
Methods:
The MobileGait workflow included video acquisition, 2D pose estimation, gait event detection, and automated computation of spatiotemporal parameters as well as sagittal-plane kinematics. For validation testing, 20 healthy elderly adults (52.6 ± 15.1 years; 13 women and 7 men) performed 5-m walk tests that were recorded simultaneously using the MobileGait iPad application and a Vicon motion-capture system. Agreement of spatiotemporal parameters with the Vicon system was assessed using intraclass correlation coefficients (ICC) and mean percentage error (MPE), while sagittal-plane kinematics were evaluated using coefficient of multiple correlation (CMC) and mean absolute error (MAE).
Results:
MobileGait demonstrated strong agreement with Vicon for step time, stride time, and cadence (ICC = 0.95-0.99; MPE <1.08%), and for knee flexion-extension (MAE ≈5.5°, CMC ≈0.96). Spatial parameters tended to be slightly overestimated (MPE <10%), while event-dependent metrics such as double-support time and ankle motion exhibited slightly greater variability.
Significance:
MobileGait enables low-cost, automated gait assessment on mobile devices, with analyses completed in minutes. It provides a practical solution to facilitate clinical gait assessment and remote monitoring in resource-limited settings, such as community centers and nursing homes.
