We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Identity and Posture Recognition in Smart Beds with Deep Multitask Learning

Abstract: Sleep posture analysis is widely used for clinical patient monitoring and sleep studies. Earlier research has revealed that sleep posture highly influences symptoms of diseases such as apnea and pressure ulcers. In this study, we propose a robust deep learning model capable of accurately detecting subjects and their sleeping postures using the publicly available data acquired from a commercial pressure mapping system. A combination of loss functions is used to discriminate subjects and their sleeping postures simultaneously. The experimental results show that our proposed method can identify the patients and their in-bed posture with almost no errors in a 10-fold cross-validation scheme. Furthermore, we show that our network achieves an average accuracy of up to 99% when faced with new subjects in a leave-one-subject-out validation procedure on the three most common sleeping posture categories. We demonstrate the effects of the combined cost function over its parameter and show that learning both tasks simultaneously improves performance significantly. Finally, we evaluate our proposed pipeline by testing it over augmented images of our dataset. The proposed algorithm can ultimately be used in clinical and smart home environments as a complementary tool with other available automated patient monitoring systems.
Comments: \c{opyright} 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works
Subjects: Machine Learning (cs.LG)
Journal reference: 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC)
DOI: 10.1109/SMC.2019.8914459
Cite as: arXiv:2104.02159 [cs.LG]
  (or arXiv:2104.02159v1 [cs.LG] for this version)

Submission history

From: Vandad Davoodnia [view email]
[v1] Mon, 5 Apr 2021 21:21:54 GMT (528kb,D)

Link back to: arXiv, form interface, contact.