Markerless Motion Capture Validated for Postural Control Research

Summary

A new peer-reviewed study confirms that Theia3D delivers highly comparable results to marker-based motion capture (Vicon) when analyzing postural control through principal component analysis (PCA).

Why Postural Control Matters

Postural movements, like quiet standing or subtle balance adjustments, are critical for understanding human balance, neuromuscular coordination, and sensorimotor integration. Traditionally, researchers rely on marker-based motion capture or force plates, which provide accurate data with some limitations:

Complex marker setup and precise placement
Errors from marker occlusion and soft tissue artifacts
Artificial constraints on movement that reduce ecological validity

Theia’s AI-driven, markerless motion capture software Theia3D overcomes these challenges by extracting whole-body pose directly from video. This enables faster, more natural, and more scalable balance assessments for both lab and real-world settings.

How the Study Was Conducted
‍

Researchers at the University of Innsbruck tested 13 participants in both eyes-open and eyes-closed quiet stance trials, capturing data with Theia3D and Vicon simultaneously. They used principal component analysis (PCA) to break down posture into its core movement strategies, called principal movements (PMs).

Key Findings
‍

Near-perfect correlation for dominant strategies: The first two principal movements, ankle-driven anterior-posterior sway and medio-lateral sway had correlations of r=0.998 and r=0.990, respectively, between Theia3D and Vicon.

Strong agreement across main postural movements: The first seven PMs (covering ~97% of total variance) showed high correlation (r=0.814-0.928), despite fundamental differences in how the systems track motion.

Sensitivity to visual input confirmed: Both systems detected increased sway speed when participants closed their eyes, aligning with decades of balance research.

Higher noise in higher-order principal movements: Theia3D showed significantly more noise in subtle postural adjustments, which can be mitigated through filtering before PCA.
‍

Why This Matters for Researchers

These results confirm that Theia3D can be trusted as a valid technology for postural control research, accurately capturing the subtle movements used by humans to maintain balance. By removing markers, researchers gain:

Faster setup and reduced participant burden
Naturalistic movement capture in labs, clinics, or field environments
Scalable solutions for large populations or multi-site studies

Even in the rigorous context of quiet standing, a condition with subtle, low-amplitude movements, Theia3D delivered results on par with gold standard.
‍

Limitations and Next Steps

This study focused on quiet stance, which is a challenging but controlled task. Future research should extend validation to dynamic postural tasks such as perturbed stance or gait. Additionally, preprocessing techniques such as filtering (available within Theia3D) should be utilized to reduce noise in higher-order PCA outputs.
‍

Conclusion

This landmark validation demonstrates that Theia3D provides highly comparable results to traditional marker-based systems (Vicon) for analyzing postural control, especially for primary balance strategies. As markerless motion capture continues to evolve, it offers biomechanics researchers an efficient, non-invasive and scalable alternative to traditional systems.
‍

Interested in applying markerless motion capture to your research?
‍

‍