Marker-Less Motion Capture Underestimates Foot Strike Angle During Walk and Run Tasks
Faculty Mentor Information
Dr. Tyler Brown, Boise State University
Additional Funding Sources
We would like to thank NIH NIGMS (2U54GM104944, P20GM109095, P20GM148321) and IFITS for their support.
Presentation Date
7-15-2026
Abstract
A larger foot strike angle (FSA, shift towards rearfoot strike) during locomotion may increase risk of lower-limb musculoskeletal injury risk by resulting in larger, faster vertical ground reaction forces. Recent advances in low-cost, portable marker-less motion capture have shown promise for accurate kinematic analysis of locomotion. Yet it is unknown if marker-less motion capture can accurately quantify FSA during common walk and run tasks. This study sought to be the first to quantify FSA with marker-less motion capture, and compare accuracy of marker-less derived FSA with FSA derived with marker based motion capture (gold standard) during a battery of walk and run tasks. 18 participants had FSA quantified using marker-less and marker-based motion capture as they walked (1.5 m/s) and ran (4.5 m/s) with and without body borne load (15 and 0 kg) over both flat and uneven surfaces. Although we were able to successfully quantify FSA with marker-less motion capture, marker-less FSA exhibited trivial to moderate relation with marker-based FSA (r: 0.03 to 0.51). Marker-less motion capture consistently underestimated FSA during both walk (MAE: ~8, nRMSE: ~ 55%) and run (MAE: ~ 12, nRMSE: ~ 100%) tasks.
Marker-Less Motion Capture Underestimates Foot Strike Angle During Walk and Run Tasks
A larger foot strike angle (FSA, shift towards rearfoot strike) during locomotion may increase risk of lower-limb musculoskeletal injury risk by resulting in larger, faster vertical ground reaction forces. Recent advances in low-cost, portable marker-less motion capture have shown promise for accurate kinematic analysis of locomotion. Yet it is unknown if marker-less motion capture can accurately quantify FSA during common walk and run tasks. This study sought to be the first to quantify FSA with marker-less motion capture, and compare accuracy of marker-less derived FSA with FSA derived with marker based motion capture (gold standard) during a battery of walk and run tasks. 18 participants had FSA quantified using marker-less and marker-based motion capture as they walked (1.5 m/s) and ran (4.5 m/s) with and without body borne load (15 and 0 kg) over both flat and uneven surfaces. Although we were able to successfully quantify FSA with marker-less motion capture, marker-less FSA exhibited trivial to moderate relation with marker-based FSA (r: 0.03 to 0.51). Marker-less motion capture consistently underestimated FSA during both walk (MAE: ~8, nRMSE: ~ 55%) and run (MAE: ~ 12, nRMSE: ~ 100%) tasks.
Comments
This poster was also presented at the 2026 Undergraduate Research Showcase. A copy can be viewed in the related record.