Markerless Knee Joint Position Measurement Using Depth Data during Stair Walking.

Area of Science:

• Biomechanics
• Medical Technology
• Rehabilitation Engineering

Background:

• Stair climbing and descending are crucial daily activities.
• Monitoring stair walking can facilitate early detection of musculoskeletal diseases.
• Markerless systems are needed for unobtrusive gait analysis during stair locomotion.

Purpose of the Study:

• To evaluate the accuracy of Microsoft Kinect v2's skeleton tracking for stair walking.
• To develop and assess a novel method using Kinect v2 depth data for 3D knee joint position estimation during stair walking.
• To compare the accuracy of the novel depth data method against Kinect v2's skeleton tracking.

Main Methods:

• Utilized Microsoft Kinect v2 for markerless motion capture during stair walking.
• Developed a new algorithm to estimate the 3D knee joint position using depth data from Kinect v2.
• Employed a 3D motion capture system as the gold standard for simultaneous measurement and comparison.

Main Results:

• The novel depth data method demonstrated higher accuracy in estimating 3D knee joint position compared to Kinect v2's skeleton tracking.
• The mean error for the depth data method was 43.2 ± 27.5 mm.
• The mean error for Kinect v2's skeleton tracking was 50.4 ± 23.9 mm.

Conclusions:

• The developed depth data method offers a more accurate approach for monitoring stair walking using Kinect v2.
• This technique holds potential for non-invasive, early detection of musculoskeletal disorders through stair activity analysis.
• Markerless monitoring of stair walking can be advanced with depth data processing for clinical applications.