Validity and reliability of kinect for measuring shoulder joint angles

Virtual reality-based physical rehabilitation, or virtual rehabilitation, provides several advantages over conventional therapy. These include the capacity to provide patient-specific treatment that adapts with functional improvements over practice, obtain quantitative measures of progress, deliver real-time performance feedback through varying modalities, and improve adherence by heightening patient motivation and entertainment. By exploiting commercially available gaming technology, virtual rehabilitation systems can even be developed at a low cost and conveniently used in the home for outpatient therapy. The Microsoft Kinect is one such gaming technology that has gained recent popularity within the virtual rehabilitation community. This highly advanced, and yet low cost, sensor enables users to interact with the system by monitoring 3D body movements, making its clinical utility highly attractive to the rehabilitation community. Before the use of this technology can be translated into the clinical setting, ensuring its precision and accuracy of measuring joint motion is paramount. The present study aimed to test the reliability and validity of upper extremity joint angle measurements with the Kinect for shoulder rehabilitation. Results indicate that while the Kinect is reliable for measuring shoulder joint angles in the frontal view, it is only valid for non-occlusive poses compared to the gold (magnetic tracker) and clinical (goniometer) standards for the shoulder.