Home // EMERGING 2015, The Seventh International Conference on Emerging Networks and Systems Intelligence // View article

Rehabilitation System in 3D Natural Scenes

Authors:
Amin Safaei
Q. M. Jonathan Wu

Keywords: Machine Vision; SoftKinetic; Motion Recognition; Video Processing; Rehabilitation

Abstract:
In this paper, we present a rehabilitation system for patients who suffer wrist injury. The idea to use computer vision for rehabilitation is not new; however, the method proposed in this paper differs significantly from previously proposed methods. We propose a 3D hand model evaluation method that can recognize soft and elaborate representations of hand motions. In practice, hand motion recognition in an unconstrained environment is a difficult task because of intra-class variation. It becomes more challenging when we lose depth data because of projection. However, the emergence of commercial depth sensors, such as Microsoft Kinect and SoftKinect, has overcome this issue. In previous work, we used the data of tip and joints, which was sufficient for simple motion; however, in complex motion, such as grabbing and rotation, it is not possible to track and estimate the depth of tips and joints. In this work, we modify the algorithm that is proposed by Rodriguez et al and Hadfield. Instead of using 2D data, we extend the method for 3D data, and for elevation information, Hidden Markov Model (HMM) is used.

Pages: 42 to 43

Copyright: Copyright (c) IARIA, 2015

Publication date: July 19, 2015

Published in: conference

ISSN: 2326-9383

ISBN: 978-1-61208-422-0

Location: Nice, France

Dates: from July 19, 2015 to July 24, 2015