Assessment of spatiotemporal gait parameters using a deep learning algorithm-based markerless motion capture system

Area of Science:

• Biomechanics
• Human Movement Analysis
• Wearable Technology

Background:

• Spatiotemporal gait parameters are crucial for assessing health status and detecting gait changes.
• Markerless motion capture offers a user-friendly, cost-effective alternative to traditional gait analysis systems.
• Reducing barriers to gait measurement can expand its clinical and research applications.

Purpose of the Study:

• To validate markerless motion capture for measuring spatiotemporal gait parameters.
• To compare markerless system measurements against marker-based motion capture and a gait mat.
• To assess the concurrent validity of markerless gait analysis.

Main Methods:

• Two studies were conducted with healthy young adults performing treadmill and over-ground gait.
• Markerless video data were synchronized with marker-based motion capture and a pressure-sensitive gait mat.
• Gait cycles were identified using kinematic heel-strike and toe-off events; nine spatiotemporal parameters were compared.

Main Results:

• Markerless motion capture demonstrated good to excellent agreement with marker-based systems and gait mats for most parameters.
• Stance time and double limb support time showed less agreement with both comparison systems.
• Stride width agreement was lower when compared to the gait mat system.

Conclusions:

• Markerless motion capture is a valid tool for measuring spatiotemporal gait parameters in healthy young adults.
• The technology shows promise for clinical and diverse settings, despite minor discrepancies in specific parameters.
• This approach can enhance accessibility to gait analysis for health assessment.