Please use this identifier to cite or link to this item: http://cmuir.cmu.ac.th/jspui/handle/6653943832/65458
Title: Kinematic-based knee angle correction for gait analysis using single kinect sensor
Authors: Tamonwan Sukumpee
Nipon Theera-Umpon
Samatchai Chamnongkich
Sansanee Auephanwiriyakul
Keywords: Chemical Engineering
Computer Science
Engineering
Mathematics
Issue Date: 8-Apr-2019
Abstract: © 2018 IEEE. Gait analysis plays an important role of offering an insight in quality of human ambulation. It is able to provide both spatial and temporal information. Furthermore, the biomechanics knowledge including kinematics and kinetics information is also extracted. Unusual walking can indicate some health problems. Therefore, it is important to consider the altering in individual's gait parameters in order to prevent an adverse situation such as falling. Besides, gait analysis can be a crucial tool of guiding the proper rehabilitation. For decades, the technologies and methods used in gait analysis have been developed. The existing technologies have found a variety of limitations whether controlled environment, extremely expensive or complicated computation need. Many research works have employed the low-cost video camera called Kinect device to the medical filed. It shows an excellent result compared to the gold standard. However, there are some problems of the accuracy needed to be improved. In this paper, we propose a novel algorithm to correct kinematics knee joint trajectory using kinematic-based algorithm to detect outliers and then each of 4 interpolation methods, i.e., Fourier series, Spline interpolation, Gaussian, and Polynomial interpolation is employed to reconstruct new data. By using Fourier series, the promising results are achieved compared to the expert measurements with the root mean square error of 4.48 degrees on average.
URI: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85065043971&origin=inward
http://cmuir.cmu.ac.th/jspui/handle/6653943832/65458
Appears in Collections:CMUL: Journal Articles

Files in This Item:
There are no files associated with this item.


Items in CMUIR are protected by copyright, with all rights reserved, unless otherwise indicated.