| 제목 |
A novel device to reconstruct 3D posture using RGB depth camera compared to EOS 3D X-ray imaging |
| 소속 |
Gangnam Severance Hospital, Department of Rehabilitation Medicine1, Yonsei University College of Medicine, Rehabilitation Institute of Neuromuscular Disease2 |
| 저자 |
Hyo Jeong Lee1,2*, Myungsang Kim1,2, Seok Young Chung1,2, Han-Eol Cho1,2, Jung Hyun Park1,2† |
Introduction
Imaging for detection of posture is important for determining cause of musculoskeletal pain. The Nobel-Prize-winning 3D EOS™ imaging system enables the acquisition of 3D reconstruction images and has extremely low radiation exposure. However, this method is asked to unnatural posture in confined space.
In this study, we propose a new device to evaluate human posture using a low-cost RGB depth camera. Human posture analyzer based on RGB depth camera (3D PA-RGBD) allows quick analysis of the user whole posture and multiple repetitive shots without exposure to radiation as well as 3D reconstruction.
The purpose of this study is to evaluate the reliability of 3D PA-RGBD and the validity compared to 3D EOS™.
Methods
A prospective study carried out in February and June 2020 included patients with musculoskeletal pain in clinic of the department of rehabilitation medicine in a tertiary hospital. The subjects who decided to use 3D EOS™ to investigate posture problems as the cause of the musculoskeletal pain were registered and were written consent.
The subjects were continuously assessed two times for 2 minutes by the 3D PA-RGBD (Moti physio, MG solutions, Seoul, Korea) with an experienced doctor to find the intra-rater reliability. Each participant stood in front of and side to the camera. Measurements were divided by standing plane; (1) coronal view (asymmetric clavicle height, pelvic oblique, bilateral Q-angles of knee, and center of 7th cervical vertebra-central sacral line (C7-CSL); (2) sagittal view (forward head posture).
Data of coronal and sagittal view parameters obtained by the 3D PA-RGBD and 3D EOS™ were compared to validate 3D PA-RGBD. The paired values of EOS imaging were also measured by another skilled physician.
Results
3.1 Subjects
A total 100 patients were recruited in the study. The mean age was 46.1 ± 17.3 (mean ± SD) and the mean body mass index was 23.3 ± 3.9.
3.2 Intra-rater reliability of 3D PA-RGBD
All intra-rater correlation coefficient for coronal and sagittal view parameters were above 0.69 and the highest parameter of the 3D PA-RGBD was C7-CSL [intra-rater correlation coefficient = 0.84] (Table 1).
3.3 Validation of 3D PA-RGBD compared to 3D EOS™
Of the parameters, C7-CSL showed that was moderate positive correlation with adjusted age and body mass index; forward head posture, asymmetric clavicle height and pelvic oblique compared to 3D EOS™ were weak to moderate positive correlation (Table 2). However, there was no significant correlation for bilateral Q-angles of knee.
Conclusion
This study presents a novel device to evaluate human posture based on RGB depth camera. In addition, the device provides excellent intra-rater reliability and moderate validation compared to EOS™ imaging system. The method with no radiation and low-cost can be an alternative to conventional radiological imaging technique to assess and evaluate posture in patients with musculoskeletal pain. |
|
Table 1. Intra-rater reliability of human posture analyzer based on RGB depth camera
Table 2. The partial correlation coefficient value for validity between human posture analyzer based on RGB depth camera and 3D EOS™
Figure 1. (A) The 3D EOS™ and (B) 3D PA-RGBD data in the same subject
|
