Agreement between a markerless and a marker-based motion capture systems for balance related quantities

Area of Science:

• Biomechanics and Movement Analysis
• Motion Capture Technology
• Human Balance and Gait Studies

Background:

• Traditional marker-based motion capture for balance studies is labor-intensive and can influence natural movement.
• Markerless motion capture technology presents a promising solution to overcome the limitations of marker-based systems.

Purpose of the Study:

• To compare dynamic balance parameters derived from marker-based and markerless motion capture systems.
• To evaluate the accuracy and reliability of markerless pose estimation software in capturing human movement for balance analysis.

Main Methods:

• Sixteen healthy participants performed four distinct motor tasks (walking, balance loss, narrow beam walking, countermovement jumps).
• Simultaneous data acquisition using marker-based and markerless (Theia3D) motion capture systems.
• Calculation of center of mass (CoM), extrapolated center of mass (XCoM), and whole-body angular momentum (WBAM); statistical comparison using Bland-Altman analysis, RMSD, and correlation.

Main Results:

• Markerless systems demonstrated minimal bias (mm range for CoM/XCoM) and low root mean square difference (RMSD) (approx. 1 cm for CoM/XCoM).
• Whole-body angular momentum (WBAM) showed less than 10% RMSD and 1% bias across directions.
• Jumping tasks exhibited the largest differences, though the source of error (markerless vs. marker-based) remains unclear.

Conclusions:

• Markerless motion capture systems provide comparable outcomes to marker-based systems for many balance study metrics.
• The findings support the use of markerless technology as a practical alternative, reducing setup time and participant interference.
• Caution is advised when analyzing highly dynamic movements like jumps due to observed larger discrepancies.