Validity and Reliability of Head Posture Measurement using Microsoft Kinect

Abstract

Introduction: Accurate and effective measurement of anomalous head posture is critical for evaluating severity of disease and determining a surgical plan for patients with ocular torticollis. For increasing the precision of head-posture measurements, several systems that use digital technologies have been proposed. To date, the common disadvantages of all such head-posture-measurement systems is that subjects are required to wear a head-mounted device and the difficult setup. To address this problem, we need a device that is contact-free and easy-to-setup. In this study, we investigated the possibility of adapting a controller-free 3D camera (Microsoft Kinect) for evaluating anomalous head posture.

Methods: Considering the cervical range of motion (CROM) as a reference, one- and three-dimensional (1D and 3D) head postures of 12 normal subjects (28–58 years of age

6 women and 6 men) were obtained using the KHT. The KHT was validated by the Pearson correlation coefficient and the intraclass correlation (ICC) coefficient. Test-retest reliability of the KHT was determined by its 95% limit of agreement (LoA) with the Bland–Altman plot. Face recognition success rate was evaluated for each head posture.

Results: The measurements of 1D and 3D head posture performed using the KHT were very close to those of the CROM with correlation coefficients of 0.99 and 0.97 (P < 0.05), respectively, as well as with an ICC of >0.99 and 0.98, respectively. The reliability tests of the KHT in terms of 1D and 3D head postures had 95% LoA angles of approximately ±2.5°and ±6.5°, respectively.

Conclusions: The KHT showed good agreement with the CROM and relatively good test–retest reliability. Considering its high performance, convenience, and low cost, KHT could be clinically used as a head-posture-measuring system.