Optimizing camera placement in multi-view markerless motion capture: a validation study against gold-standard...

Area of Science:

• Biomechanics
• Motion Capture Technology
• Ergonomics

Background:

• Markerless motion capture offers a cost-effective alternative to traditional laboratory systems.
• Understanding the impact of camera configuration on measurement validity is crucial for system deployment.
• Existing research often lacks detailed validation of camera arrangements for markerless systems.

Purpose of the Study:

• To determine the optimal camera configuration for a markerless motion capture system.
• To validate the system's performance against a gold standard in static and dynamic tasks.
• To provide practical guidance for affordable motion capture system setup.

Main Methods:

• Six stereo camera pairs were calibrated for 3D reconstruction using MediaPipe.
• Joint angles were computed from markerless motion capture data.
• Agreement with a gold standard was assessed using Mean Absolute Error (MAE), Intraclass Correlation Coefficient (ICC), and Bland-Altman analysis.

Main Results:

• System performance was highly sensitive to camera geometry; optimal configuration was ~45 degrees separation.
• Excellent agreement was achieved for static (ICC=0.851, MAE=9.28°) and dynamic (ICC=0.823, MAE=12.92°) tasks.
• Major joints showed good to excellent reliability (ICC=0.73-0.92), while neck agreement was moderate (ICC=0.56-0.66).

Conclusions:

• A two-camera setup with ~45-degree separation can yield clinically useful joint angle estimates.
• Markerless motion capture systems are sensitive to camera geometry, necessitating careful configuration.
• This study provides practical guidance for deploying affordable motion capture in biomechanics, ergonomics, and rehabilitation.