Biomechanical Reconstruction with Confidence Intervals from Multiview Markerless Motion Capture.

Area of Science:

• Biomechanics
• Motion Capture Technology
• Clinical Movement Analysis

Background:

• Multiview markerless motion capture (MMMC) offers high-quality movement analysis potential.
• Existing validation studies provide average performance but lack individual-specific confidence intervals crucial for clinical use.
• There is a need for methods providing confidence intervals for specific kinematic estimates in MMMC.

Purpose of the Study:

• To extend previous work by developing a method to estimate confidence intervals for individual kinematic estimates in MMMC.
• To provide clinicians and researchers with reliable uncertainty measures for motion capture data.
• To enable identification of trials with high kinematic uncertainty.

Main Methods:

• Utilized an implicit representation of trajectories optimized end-to-end through a differentiable biomechanical model.
• Employed variational approximation to learn the posterior probability distribution over pose given detected keypoints.
• Estimated confidence intervals for individual joints and joint angles over time.

Main Results:

• Confidence intervals for virtual marker locations were generally within 10-15 mm spatial error.
• Confidence intervals for joint angles were typically a few degrees, widening for distal joints.
• The method successfully modeled correlations between joint angles (e.g., hip and pelvis).

Conclusions:

• The developed method provides reliable, individual-specific confidence intervals for MMMC kinematic data.
• These confidence intervals are essential for assessing the reliability of motion analysis in clinical and research settings.
• The ability to identify high kinematic uncertainty improves the trustworthiness and applicability of MMMC.