Postural control assessment via Microsoft Azure Kinect DK: An evaluation study

Area of Science:

• Biomechanics and Human Movement Analysis
• Rehabilitation Engineering
• Wearable Technology and Sensor Systems

Background:

• Consumer-grade motion tracking systems like Microsoft Kinect have been adapted for postural control assessment.
• The new Azure Kinect DK, designed for professional applications, features improved hardware but requires accuracy validation for postural control.
• Existing systems offer non-invasive, low-cost tracking suitable for home rehabilitation, but professional-grade accuracy is key.

Purpose of the Study:

• To evaluate the accuracy of the Azure Kinect DK for postural control assessment.
• To compare Azure Kinect DK motion tracking data against a Vicon 3D system, the gold standard for motion capture.
• To determine the suitability of the Azure Kinect DK for home-based rehabilitation applications.

Main Methods:

• A comparative study was conducted involving 26 participants.
• Subjects performed functional reach (lateral and forward) and functional balance exercises.
• Motion data from the Azure Kinect DK was compared against a Vicon 3D motion tracking system.

Main Results:

• Azure Kinect DK demonstrated highly accurate tracking of major body joints during lateral reach movements.
• Root Mean Square Error (RMSE) between systems was approximately 0.2 for reach exercises and 0.47 for balance exercises.
• Angular Mean Absolute Error ranged from 5-15 degrees for upper body joints, with higher errors in lower body joints, especially during faster movements.

Conclusions:

• The Azure Kinect DK exhibits significant potential for use in home rehabilitation settings.
• Accurate postural control assessment is a critical component of effective rehabilitation.
• The system's accuracy supports its application in developing advanced home rehabilitation solutions.