Concurrent Validity of a Custom Method for Markerless 3D Full-Body Motion Tracking of Children and Young Adults Based on a Single RGB-D Camera

Abstract

Low-cost, portable RGB-D cameras with integrated body tracking functionality enable easy-to-use 3D motion analysis without requiring expensive facilities and specialized personnel. However, the accuracy of existing systems is insufficient for most clinical applications. In this study, we investigated the concurrent validity of our custom tracking method based on RGB-D images with respect to a gold-standard marker-based system. Additionally, we analyzed the validity of the publicly available Microsoft Azure Kinect Body Tracking (K4ABT). We recorded 23 typically developing children and healthy young adults (aged 5 to 29 years) performing five different movement tasks using a Microsoft Azure Kinect RGB-D camera and a marker-based multi-camera Vicon system simultaneously. Our method achieved a mean per joint position error over all joints of 11.7 mm compared to the Vicon system, and 98.4% of the estimated joint positions had an error of less than 50 mm. Pearson’s correlation coefficients ${r}$ ranged from strong ( ${r}$ =0.64) to almost perfect ( ${r}>$ 0.99). K4ABT demonstrated satisfactory accuracy most of the time but showed short periods of tracking failures in nearly two-thirds of all sequences limiting its use for clinical motion analysis. In conclusion, our tracking method highly agrees with the gold standard system. It paves the way towards a low-cost, easy-to-use, portable 3D motion analysis system for children and young adults.

Keywords

• Children
• Kinect
• 3D motion tracking
• RGB-D
• Vicon