Artifacts' roaming beats recognition for estimating care activities in a nursing home

Francisco E. Martinez-Perez; J. Angel Gonzalez-Fraga; Monica Tentori

4th International ICST Conference on Pervasive Computing Technologies for Healthcare

Research Article

Artifacts' roaming beats recognition for estimating care activities in a nursing home

Download690 downloads

Cite: BibTeX Plain Text

@INPROCEEDINGS{10.4108/ICST.PERVASIVEHEALTH2010.8847,
    author={Francisco E. Martinez-Perez and J. Angel Gonzalez-Fraga and Monica Tentori},
    title={Artifacts' roaming beats recognition for estimating care activities in a nursing home},
    proceedings={4th International ICST Conference on Pervasive Computing Technologies for Healthcare},
    proceedings_a={PERVASIVEHEALTH},
    year={2010},
    month={6},
    keywords={Activity estimation context-aware computing object recognition correiationjiiters artifacts' roaming beats},
    doi={10.4108/ICST.PERVASIVEHEALTH2010.8847}
}

Francisco E. Martinez-Perez
J. Angel Gonzalez-Fraga
Monica Tentori
Year: 2010
Artifacts' roaming beats recognition for estimating care activities in a nursing home
PERVASIVEHEALTH
ICST
DOI: 10.4108/ICST.PERVASIVEHEALTH2010.8847

Francisco E. Martinez-Perez¹^,*, J. Angel Gonzalez-Fraga²^,*, Monica Tentori^3,4^,*

1: Facultad de Ingenieria, Universidad Autonoma de Baja, California, Ensenada, Mexico
2: Facultad de Ciencias, Universidad Autonoma de Baja, California, Ensenada, Mexico
3: Department of Informatics, UCI, Irvine, USA.
4: Facultad de Ciencias, UABC, Ensenada, Mexico

*Contact email: fmartinezperez@acm.org, angel_fraga@uabc.edu.mx, mtentori@uabc.mx

Abstract

This paper describes an approach to estimate the activity being performed by caregivers in a nursing home. The approach is based on information gathered from a workplace study conducted in a nursing home, in which 400 hours of detailed observation caregivers' activities were recorded. Automatic identification of objects based in composite correlation filters was implemented in order to generate a descriptive behavior of objects. When the behavior of the artifacts related with the activities is validated, the estimation of an activity is found. The activities estimated with a 92.72% of accuracy (on average) include hygiene, feeding and taking vital signs.

Keywords: Activity estimation context-aware computing object recognition correiationjiiters artifacts' roaming beats

Published: 2010-06-07

: http://dx.doi.org/10.4108/ICST.PERVASIVEHEALTH2010.8847