This paper proposes an online, data-driven method to detect in which position (lying or standing) a women is performing Kegel exercises from measurements collected with a vaginal pressure sensor array. Pressure data has been collected with the vaginal pressure sensor from women performing Kegel exercises by playing a dedicated mobile app, which is controlled by contracting their pelvic floor muscles. Depending on their position while playing (lying or standing), the recorded pressure patterns exhibit different characteristics in terms of intensity, location and width of the pressure peak, which may be used to detect the human position. For this, the recorded data is filtered, opportune features are extracted and a suitable classifier is trained to distinguish the two positions. The results show that the human position can be accurately detected online when using individual models for each patient (in our experiments, up to 1% of false positives and 4% false negatives), whereas the detection capabilities might decrease drastically when considering the same classifier for another women (e.g., up to 95% of false positives).

Online, data-driven detection of human position during Kegel exercising

Varagnolo D.;
2020

Abstract

This paper proposes an online, data-driven method to detect in which position (lying or standing) a women is performing Kegel exercises from measurements collected with a vaginal pressure sensor array. Pressure data has been collected with the vaginal pressure sensor from women performing Kegel exercises by playing a dedicated mobile app, which is controlled by contracting their pelvic floor muscles. Depending on their position while playing (lying or standing), the recorded pressure patterns exhibit different characteristics in terms of intensity, location and width of the pressure peak, which may be used to detect the human position. For this, the recorded data is filtered, opportune features are extracted and a suitable classifier is trained to distinguish the two positions. The results show that the human position can be accurately detected online when using individual models for each patient (in our experiments, up to 1% of false positives and 4% false negatives), whereas the detection capabilities might decrease drastically when considering the same classifier for another women (e.g., up to 95% of false positives).
2020
IFAC-PapersOnLine
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11577/3495161
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
social impact