Multi-sensors Analysis for Everyday Activity Monitoring

LES AUTEURS Nadia Zouba, François Bremond, Monique Thonnat, Van Thinh Vu

RESUME:
In this paper, we present a multimodal (video and contact sensors) monitoring approach for the automatic monitoring everyday activities of elderly people. The specificity of our approach consists in the integration of video analysis with other information resulting from contact sensors installed at home in order to improve the recognition of observed activities. After reviewing the state of the art, we describe the modeling of the scenarios of interest specified by medical experts. Then we present the multimodal event recognition process based on video and contact sensor analysis. Finally, we conclude by presenting preliminary results.

Mots clé: event recognition, scene understanding, multimodal behavior analysis.

Pour télécharger cet article,  cliquez ici

BibTeX reference:

@InProceedings{Zouba_SETIT_2007,
author =       {N. Zouba and F. Bremond and M. Thonnat and V. T. Vu},
title =        {Multi-sensors Analysis for Everyday Elderly Activity 
Monitoring},
booktitle =    {Proceedings of SETIT 2007},
month =        {March}
year =         {2007}
address =      {Tunis, TUNISIA}
}