Advanced search
2 files | 1.05 MB Add to list

Tailoring radar-based patient monitoring models to real-life needs using utility maximization

Author
Organization
Project
Abstract
Monitoring patients in hospitals or care homes using radars is an interesting problem with life-saving applications. Today deep neural networks are employed for patient monitoring, but these do not provide uncertainties, nor do they consider the asymmetry in the real life cost of misclassifying different activities. In this work we use Bayesian Neural Networks that provide uncertainty on their predictions. We combine these models with a self-defined utility function to obtain tailored predictions that are more conservative for classes where misclassifications come at a higher risk or cost. We show that Bayesian neural networks are more robust, and generalize better on radar human-activity images than deterministic ones, and that they are able to reduce the cost of misclassifications in a realistic example setting by 37% compared to approaches from literature.
Keywords
CLASSIFICATION, Bayesian Deep Learning, Radar Sensors, Human Activity Recognition, Patient Monitoring, Trustworthy, Fall Detection

Downloads

  • (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 535.35 KB
  • conference paper.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 518.69 KB

Citation

Please use this url to cite or link to this publication:

MLA
Vincent-De Sloover, Louis, et al. “Tailoring Radar-Based Patient Monitoring Models to Real-Life Needs Using Utility Maximization.” 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), IEEE, 2022, pp. 81–84.
APA
Vincent-De Sloover, L., Werthen-Brabants, L., Bhavanasi, G., Dhaene, T., & Couckuyt, I. (2022). Tailoring radar-based patient monitoring models to real-life needs using utility maximization. 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), 81–84. NEW YORK: IEEE.
Chicago author-date
Vincent-De Sloover, Louis, Lorin Werthen-Brabants, Geethika Bhavanasi, Tom Dhaene, and Ivo Couckuyt. 2022. “Tailoring Radar-Based Patient Monitoring Models to Real-Life Needs Using Utility Maximization.” In 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), 81–84. NEW YORK: IEEE.
Chicago author-date (all authors)
Vincent-De Sloover, Louis, Lorin Werthen-Brabants, Geethika Bhavanasi, Tom Dhaene, and Ivo Couckuyt. 2022. “Tailoring Radar-Based Patient Monitoring Models to Real-Life Needs Using Utility Maximization.” In 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), 81–84. NEW YORK: IEEE.
Vancouver
1.
Vincent-De Sloover L, Werthen-Brabants L, Bhavanasi G, Dhaene T, Couckuyt I. Tailoring radar-based patient monitoring models to real-life needs using utility maximization. In: 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD). NEW YORK: IEEE; 2022. p. 81–4.
IEEE
[1]
L. Vincent-De Sloover, L. Werthen-Brabants, G. Bhavanasi, T. Dhaene, and I. Couckuyt, “Tailoring radar-based patient monitoring models to real-life needs using utility maximization,” in 2022 19TH EUROPEAN RADAR CONFERENCE (EURAD), Milan, ITALY, 2022, pp. 81–84.
@inproceedings{01GVD816VSFM0HBE5PB4MSEK56,
  abstract     = {{Monitoring patients in hospitals or care homes using radars is an interesting problem with life-saving applications. Today deep neural networks are employed for patient monitoring, but these do not provide uncertainties, nor do they consider the asymmetry in the real life cost of misclassifying different activities. In this work we use Bayesian Neural Networks that provide uncertainty on their predictions. We combine these models with a self-defined utility function to obtain tailored predictions that are more conservative for classes where misclassifications come at a higher risk or cost. We show that Bayesian neural networks are more robust, and generalize better on radar human-activity images than deterministic ones, and that they are able to reduce the cost of misclassifications in a realistic example setting by 37% compared to approaches from literature.}},
  author       = {{Vincent-De Sloover, Louis and Werthen-Brabants, Lorin and Bhavanasi, Geethika and Dhaene, Tom and Couckuyt, Ivo}},
  booktitle    = {{2022 19TH EUROPEAN RADAR CONFERENCE (EURAD)}},
  isbn         = {{9782874870712}},
  keywords     = {{CLASSIFICATION,Bayesian Deep Learning,Radar Sensors,Human Activity Recognition,Patient Monitoring,Trustworthy,Fall Detection}},
  language     = {{eng}},
  location     = {{Milan, ITALY}},
  pages        = {{81--84}},
  publisher    = {{IEEE}},
  title        = {{Tailoring radar-based patient monitoring models to real-life needs using utility maximization}},
  year         = {{2022}},
}

Web of Science
Times cited: