Multi-Kinect Spinal Motion Measurement Method Based on Markings
Three-dimensional spinal motion measurement is of great significance in fields such as rehabilitation medicine, sports science, and biomechanics. At present, optical motion capture systems are abundant, among which Vicon and OptiTrack are commonly used for three-dimensional motion measurement of the spine. Although they have high accuracy, they are expensive and occupy a large space, which limits their widespread application in outpatient screening. As a low-cost, easy-to-use, and highly reliable depth sensor, Kinect has the potential for three-dimensional motion measurement of the spine. The marker based multi-Kinect spine motion measurement method introduces marker points and multi-perspective data registration techniques, aiming to solve the problem of unmarked bone recognition being unable to describe changes in spine morphology, as well as the limitation of a single camera being unable to capture six degrees of freedom motion data of the spine. The final outcomes of the experiment made a validation of the accuracy of the method in evaluating spinal morphology (average MED of 2.7mm) and six degree of freedom motion data (average RMSE error of 3.7mm). This method not only ensures measurement accuracy, but also significantly reduces cost and operational complexity, and has broad application prospects.
Authors
- Wei Guo (ORCID: https://orcid.org/0000-0002-8009-2595)
- Yuxuan Peng (ORCID: https://orcid.org/0000-0002-8448-4991)
- Fei Wang (ORCID: https://orcid.org/0000-0002-6918-3779)
- Xinyu Wu (ORCID: https://orcid.org/0000-0002-7218-238X)
- Xinfeng Zhang
- Xiaoming Liu (ORCID: https://orcid.org/0000-0002-7631-751X)
- Chun Gan
- Xiaoman Wang (ORCID: https://orcid.org/0009-0004-4929-1184)
Institutions
- Beijing University of Technology (CN)
- Air Force General Hospital PLA (CN)
Publication Details
- Published
- 2024-11-24
- DOI
- https://doi.org/10.1109/iccvit63928.2024.10872452
- Primary Topic
- Medical Imaging and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00