Clinical Whole-Body Gait Characterization Using a Single RGB-D Sensor

Area of Science:

• Biomedical Engineering
• Computer Vision
• Clinical Biomechanics

Background:

• Instrumented gait analysis is crucial for early detection of neurological disorders, disease progression monitoring, and fall risk assessment.
• Traditional marker-based 3D motion analysis is resource-intensive, limiting its widespread clinical adoption.
• Computer vision advancements enable accurate markerless human motion tracking.

Purpose of the Study:

• To present vGait, a novel 3D gait assessment method utilizing a single RGB-D sensor and advanced pose-tracking algorithms.
• To validate the accuracy and reliability of vGait in quantifying gait parameters and coordination metrics.
• To demonstrate the potential of vGait for accessible and scalable clinical and non-clinical mobility monitoring.

Main Methods:

• Development of vGait, a markerless 3D gait analysis system employing a single RGB-D sensor.
• Utilizing state-of-the-art pose-tracking algorithms for whole-body tracking.
• Validation in healthy participants during walking from frontal and sagittal perspectives.

Main Results:

• vGait demonstrated high accuracy in detecting initial and final foot contacts (F1 scores > 95%).
• Reliable quantification of spatiotemporal gait parameters (e.g., stride time, stride length) and whole-body coordination metrics (e.g., arm swing, knee ROM).
• Comparable performance across frontal and sagittal walking perspectives, with detailed granularity (mean, variability, asymmetry).

Conclusions:

• vGait provides a flexible, accurate, and resource-efficient 3D gait assessment solution.
• The system is suitable for diverse clinical and non-clinical settings, including outpatient clinics and community environments.
• vGait can enhance diagnostic/therapeutic workflows and broaden access to clinical mobility monitoring.