Verification of Markerless Gait Analysis: Multi-Camera and Single-Camera Approaches in Comparison to Marker-Based...

Area of Science:

• Biomechanics
• Human Movement Analysis
• Gait Analysis

Background:

• Marker-based gait analysis is the gold standard but is labor-intensive.
• Markerless gait analysis offers a potentially more accessible alternative.
• Validating markerless systems against marker-based systems is crucial for clinical adoption.

Purpose of the Study:

• To validate single-camera markerless (S-ML) and multi-camera markerless (M-ML) gait analysis systems.
• To compare the reliability and accuracy of S-ML and M-ML against a marker-based (MB) system.
• To assess the performance of markerless systems for spatiotemporal and kinematic gait parameters.

Main Methods:

• Sixteen healthy adults performed walking trials.
• Simultaneous gait analysis using S-ML, M-ML, and MB systems.
• Intraclass correlation coefficients and root mean squared error (RMSE) were used for analysis.

Main Results:

• Both S-ML and M-ML showed good-to-excellent reliability for spatiotemporal parameters like step length and gait speed.
• S-ML and M-ML demonstrated comparable joint kinematics, especially in the sagittal plane (hip and knee).
• S-ML showed lower RMSE for knee abduction/adduction but reduced agreement for ankle joint angles compared to MB.

Conclusions:

• Markerless gait analysis, including S-ML and M-ML, is a viable alternative to conventional marker-based methods.
• S-ML offers promising results, comparable to M-ML in several aspects.
• Further improvements are needed for specific gait parameters, particularly ankle kinematics, to enhance markerless system accuracy.