Home // International Journal On Advances in Life Sciences, volume 11, numbers 3 and 4, 2019 // View article

Evaluation of Joint Range of Motion Measured by Vision Cameras

Authors:
Oky Dicky Ardiansyah Prima
Yuta Ono
Yoshitoshi Murata
Hisayoshi Ito
Takashi Imabuchi
Yukihide Nishimura

Keywords: computer vision; range of motion; activities of daily living; 3D human pose estimation

Abstract:
Joint Range of Motion (ROM) can be measured through a variety of methods including the use of sophisticated devices such as goniometers and non-intrusive three-dimensional (3D) sensor devices such as motion capture systems. The Microsoft Kinect has been proposed as an affordable motion capture device as an alternative to goniometers. However, due to limited measurement range and complex setup, this device cannot be used as a self-measurement during home rehabilitation or flexibility training. With the recent progress in human pose estimation based on computer vision approaches, it has become possible to estimate human joint positions in real time from vision cameras. This study evaluates joint ROM measured by two vision cameras using 3D human pose estimation based on a single camera and a stereo camera. The ROM of major joints, which consist of shoulders, elbows, a hip, and knees was evaluated for 10 users. The stereo camera gives the best results with a small bias to the goniometer compared to the single camera and the Kinect. Vision cameras have advantages on estimating semi-occluded joint locations than the Kinect. The 3D human pose based on a single camera opens up possibilities to build Tele-Rehabilitation (TR).

Pages: 128 to 137

Copyright: Copyright (c) to authors, 2019. Used with permission.

Publication date: December 30, 2019

Published in: journal

ISSN: 1942-2660