Smartweb-based platform to support physical rehabilitation

Yves Rybarczyk, Jan Kleine Deters, Clément Cointe, Danilo Esparza

Research output: Contribution to journalArticle

14 Citations (Scopus)
2 Downloads (Pure)

Abstract

The enhancement of ubiquitous and pervasive computing brings new perspectives in medical rehabilitation. In that sense, the present study proposes a smart, web-based platform to promote the reeducation of patients after hip replacement surgery. This project focuses on two fundamental aspects in the development of a suitable tele-rehabilitation application, which are: (i) being based on an affordable technology, and (ii) providing the patients with a real-time assessment of the correctness of their movements. A probabilistic approach based on the development and training of ten Hidden Markov Models (HMMs) is used to discriminate in real time the main faults in the execution of the therapeutic exercises. Two experiments are designed to evaluate the precision of the algorithm for classifying movements performed in the laboratory and clinical settings, respectively. A comparative analysis of the data shows that the models are as reliable as the physiotherapists to discriminate and identify the motion errors. The results are discussed in terms of the required setup for a successful application in the field and further implementations to improve the accuracy and usability of the system.

Original languageEnglish
Article number1344
JournalSensors (Switzerland)
Volume18
Issue number5
DOIs
Publication statusPublished - 1 May 2018

    Fingerprint

Keywords

  • Hidden markov model
  • Motion assessment
  • Motor rehabilitation
  • Natural user interface
  • Telemedicine

Cite this