| Art der Veröffentlichung: |
Artikel in Konferenzband |
| Autor: |
Christian Mandel, Amit Choudhury, Serge Autexier, Karin Hochbaum, Jeannine Budelmann, Petra Wedler |
| Titel: |
Gait cycle classification for wheeled walker users by matching time series of distance measurements |
| Buch / Sammlungs-Titel: |
Proceedings of the 2018 Conference of the Rehabilitation Engineering and Assistive Technology Society of North America |
| Erscheinungsjahr: |
2018 |
| Abstract / Kurzbeschreibung: |
In this work we present a system that classifies the gait of wheeled walker users w.r.t. 14 gait properties, which are commonly used during the clinical assessment of elderly persons’ gait. By sampling the image streams of walker-mounted depth cameras with eight virtual distance measurements pointing to relevant body features, the system constructs time series of measurements that are used for training the gait cycle classifiers. Within a clinical observation study that involved 26 patients aged 47-93 years ( mean age: 74.3 years), the system showed an
average classification rate of 96.9% over all subjects and gait properties, when using the Mahalanobis distance on time series during the test phase of the classification system.
We propose the system presented as the basis for a user interface that informs people depending on wheeled walkers about insecure gait and postures, and thus has the ability to reduce the risk of fall. |
| Status: |
Reviewed |
| Letzte Aktualisierung: |
22. 06. 2018 |
|
 |
