Descripción
|
|
---|---|
In order to perform a continuous monitoring of patients in their daily lives, it is the need to contextualize the collected data coming from the patients. In this sense, the identification of the Activities of the Daily Life (ADL) carried out by the subjects is essential to understand and to put in context other data linked to the monitoring disease itself. This work was aimed at using the accelerometers integrated in most of the current available smartphones to build an automatic activity recognizer based on the signal coming from this sort of sensors. The validation carried out showed the impact of the frequency sampling in the classifier performance as well as the impact in the battery usage. Finally, it suggested that the use of one second as sampling period is a fair trade-off between accuracy in the classification and power saving. Keywords: Accelerometer, ADL, Machine Learning | |
Internacional
|
Si |
Nombre congreso
|
IUPESM World Congress on Medical Physics and Biomedical Engineering (IFBME) 2015. |
Tipo de participación
|
960 |
Lugar del congreso
|
Toronto, Canadá |
Revisores
|
Si |
ISBN o ISSN
|
978-3-319-19386-1 |
DOI
|
10.1007/978-3-319-19387 |
Fecha inicio congreso
|
07/06/2015 |
Fecha fin congreso
|
12/06/2015 |
Desde la página
|
1457 |
Hasta la página
|
1460 |
Título de las actas
|
Proceedings of the IUPESM World Congress on Medical Physics and Biomedical Engineering (IFBME) 2015. www.wc.2015.org .Volume 51 |