Self-Supervised Representation Learning and Temporal-Spectral Feature Fusion for Bed Occupancy Detection

In automated sleep monitoring systems, bed occupancy detection is the foundation or the first step before other downstream tasks, such as inferring sleep activities and vital signs. The existing methods do not generalize well to real-world environments due to single environment settings and rely on...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings of ACM on interactive, mobile, wearable and ubiquitous technologies Vol. 8; no. 3
Main Authors: Song, Yingjian, Pitafi, Zaid Farooq, Dou, Fei, Sun, Jin, Zhang, Xiang, Phillips, Bradley G, Song, Wenzhan
Format: Journal Article
Language:English
Published: United States 01.09.2024
Subjects:
ISSN:2474-9567, 2474-9567
Online Access:Get more information
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract In automated sleep monitoring systems, bed occupancy detection is the foundation or the first step before other downstream tasks, such as inferring sleep activities and vital signs. The existing methods do not generalize well to real-world environments due to single environment settings and rely on threshold-based approaches. Manually selecting thresholds requires observing a large amount of data and may not yield optimal results. In contrast, acquiring extensive labeled sensory data poses significant challenges regarding cost and time. Hence, developing models capable of generalizing across diverse environments with limited data is imperative. This paper introduces SeismoDot, which consists of a self-supervised learning module and a spectral-temporal feature fusion module for bed occupancy detection. Unlike conventional methods that require separate pre-training and fine-tuning, our self-supervised learning module is co-optimized with the primary target task, which directs learned representations toward a task-relevant embedding space while expanding the feature space. The proposed feature fusion module enables the simultaneous exploitation of temporal and spectral features, enhancing the diversity of information from both domains. By combining these techniques, SeismoDot expands the diversity of embedding space for both the temporal and spectral domains to enhance its generalizability across different environments. SeismoDot not only achieves high accuracy (98.49%) and F1 scores (98.08%) across 13 diverse environments, but it also maintains high performance (97.01% accuracy and 96.54% F1 score) even when trained with just 20% (4 days) of the total data. This demonstrates its exceptional ability to generalize across various environmental settings, even with limited data availability.
AbstractList In automated sleep monitoring systems, bed occupancy detection is the foundation or the first step before other downstream tasks, such as inferring sleep activities and vital signs. The existing methods do not generalize well to real-world environments due to single environment settings and rely on threshold-based approaches. Manually selecting thresholds requires observing a large amount of data and may not yield optimal results. In contrast, acquiring extensive labeled sensory data poses significant challenges regarding cost and time. Hence, developing models capable of generalizing across diverse environments with limited data is imperative. This paper introduces SeismoDot, which consists of a self-supervised learning module and a spectral-temporal feature fusion module for bed occupancy detection. Unlike conventional methods that require separate pre-training and fine-tuning, our self-supervised learning module is co-optimized with the primary target task, which directs learned representations toward a task-relevant embedding space while expanding the feature space. The proposed feature fusion module enables the simultaneous exploitation of temporal and spectral features, enhancing the diversity of information from both domains. By combining these techniques, SeismoDot expands the diversity of embedding space for both the temporal and spectral domains to enhance its generalizability across different environments. SeismoDot not only achieves high accuracy (98.49%) and F1 scores (98.08%) across 13 diverse environments, but it also maintains high performance (97.01% accuracy and 96.54% F1 score) even when trained with just 20% (4 days) of the total data. This demonstrates its exceptional ability to generalize across various environmental settings, even with limited data availability.In automated sleep monitoring systems, bed occupancy detection is the foundation or the first step before other downstream tasks, such as inferring sleep activities and vital signs. The existing methods do not generalize well to real-world environments due to single environment settings and rely on threshold-based approaches. Manually selecting thresholds requires observing a large amount of data and may not yield optimal results. In contrast, acquiring extensive labeled sensory data poses significant challenges regarding cost and time. Hence, developing models capable of generalizing across diverse environments with limited data is imperative. This paper introduces SeismoDot, which consists of a self-supervised learning module and a spectral-temporal feature fusion module for bed occupancy detection. Unlike conventional methods that require separate pre-training and fine-tuning, our self-supervised learning module is co-optimized with the primary target task, which directs learned representations toward a task-relevant embedding space while expanding the feature space. The proposed feature fusion module enables the simultaneous exploitation of temporal and spectral features, enhancing the diversity of information from both domains. By combining these techniques, SeismoDot expands the diversity of embedding space for both the temporal and spectral domains to enhance its generalizability across different environments. SeismoDot not only achieves high accuracy (98.49%) and F1 scores (98.08%) across 13 diverse environments, but it also maintains high performance (97.01% accuracy and 96.54% F1 score) even when trained with just 20% (4 days) of the total data. This demonstrates its exceptional ability to generalize across various environmental settings, even with limited data availability.
In automated sleep monitoring systems, bed occupancy detection is the foundation or the first step before other downstream tasks, such as inferring sleep activities and vital signs. The existing methods do not generalize well to real-world environments due to single environment settings and rely on threshold-based approaches. Manually selecting thresholds requires observing a large amount of data and may not yield optimal results. In contrast, acquiring extensive labeled sensory data poses significant challenges regarding cost and time. Hence, developing models capable of generalizing across diverse environments with limited data is imperative. This paper introduces SeismoDot, which consists of a self-supervised learning module and a spectral-temporal feature fusion module for bed occupancy detection. Unlike conventional methods that require separate pre-training and fine-tuning, our self-supervised learning module is co-optimized with the primary target task, which directs learned representations toward a task-relevant embedding space while expanding the feature space. The proposed feature fusion module enables the simultaneous exploitation of temporal and spectral features, enhancing the diversity of information from both domains. By combining these techniques, SeismoDot expands the diversity of embedding space for both the temporal and spectral domains to enhance its generalizability across different environments. SeismoDot not only achieves high accuracy (98.49%) and F1 scores (98.08%) across 13 diverse environments, but it also maintains high performance (97.01% accuracy and 96.54% F1 score) even when trained with just 20% (4 days) of the total data. This demonstrates its exceptional ability to generalize across various environmental settings, even with limited data availability.
Author Pitafi, Zaid Farooq
Sun, Jin
Song, Wenzhan
Song, Yingjian
Dou, Fei
Zhang, Xiang
Phillips, Bradley G
Author_xml – sequence: 1
  givenname: Yingjian
  surname: Song
  fullname: Song, Yingjian
  organization: The University of Georgia, USA, Athens, Georgia
– sequence: 2
  givenname: Zaid Farooq
  surname: Pitafi
  fullname: Pitafi, Zaid Farooq
  organization: The University of Georgia, USA, Athens, Georgia
– sequence: 3
  givenname: Fei
  surname: Dou
  fullname: Dou, Fei
  organization: The University of Georgia, USA, Athens, Georgia
– sequence: 4
  givenname: Jin
  surname: Sun
  fullname: Sun, Jin
  organization: The University of Georgia, USA, Athens, Georgia
– sequence: 5
  givenname: Xiang
  surname: Zhang
  fullname: Zhang, Xiang
  organization: University of North Carolina at Charlotte, USA, Athens, Georgia
– sequence: 6
  givenname: Bradley G
  surname: Phillips
  fullname: Phillips, Bradley G
  organization: The University of Georgia, USA, Athens, Georgia
– sequence: 7
  givenname: Wenzhan
  surname: Song
  fullname: Song, Wenzhan
  organization: The University of Georgia, USA, Athens, Georgia
BackLink https://www.ncbi.nlm.nih.gov/pubmed/40093941$$D View this record in MEDLINE/PubMed
BookMark eNpNkE1Lw0AYhBep2FqL_0D26CW6X-lmj35VhULB1nN4s3mjkWQTd7NC_70RKziXGZiHOcwpmbjOISHnnF1xrtJrudRZytURmQmlVWLSpZ78y1OyCOGDMcaNlBnTJ2SqGDPSKD4j71tsqmQbe_RfdcCSvmDvMaAbYKg7R9cI3tXujYIr6Q7bvvPQJNse7TAGukIYoke6iuGHrjpPb8eRjbWxB2f39B6HER27M3JcQRNwcfA5eV097O6ekvXm8fnuZp2AZGxIZKmVYkoZjTwrR4HFMjXIURdVxgUywSBdygwNSAu8AG6tFSCUKbRNhZiTy9_d3nefEcOQt3Ww2DTgsIshl1xnwgjG5YheHNBYtFjmva9b8Pv87x3xDdWnaMc
CitedBy_id crossref_primary_10_1145_3729464
ContentType Journal Article
DBID NPM
7X8
DOI 10.1145/3678514
DatabaseName PubMed
MEDLINE - Academic
DatabaseTitle PubMed
MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
PubMed
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod no_fulltext_linktorsrc
Discipline Engineering
EISSN 2474-9567
ExternalDocumentID 40093941
Genre Journal Article
GrantInformation_xml – fundername: NHLBI NIH HHS
  grantid: R01 HL172291
GroupedDBID AAKMM
AAYFX
ACM
AEFXT
AIKLT
ALMA_UNASSIGNED_HOLDINGS
GUFHI
LHSKQ
NPM
ROL
7X8
ID FETCH-LOGICAL-a300t-3d74404497e18ddddaced59e1e7bf812e020a5638e9a3ca1ba1ccc2a249b7c522
IEDL.DBID 7X8
ISICitedReferencesCount 1
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001318802900004&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 2474-9567
IngestDate Fri Jul 11 14:49:41 EDT 2025
Thu Sep 11 02:53:09 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 3
Keywords Self-Supervised Learning
Bed Occupancy
Spectrum-temporal feature fusion
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-a300t-3d74404497e18ddddaced59e1e7bf812e020a5638e9a3ca1ba1ccc2a249b7c522
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
OpenAccessLink https://dl.acm.org/doi/10.1145/3678514
PMID 40093941
PQID 3178292013
PQPubID 23479
ParticipantIDs proquest_miscellaneous_3178292013
pubmed_primary_40093941
PublicationCentury 2000
PublicationDate 2024-09-01
PublicationDateYYYYMMDD 2024-09-01
PublicationDate_xml – month: 09
  year: 2024
  text: 2024-09-01
  day: 01
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle Proceedings of ACM on interactive, mobile, wearable and ubiquitous technologies
PublicationTitleAlternate Proc ACM Interact Mob Wearable Ubiquitous Technol
PublicationYear 2024
SSID ssj0001933807
Score 2.2737472
Snippet In automated sleep monitoring systems, bed occupancy detection is the foundation or the first step before other downstream tasks, such as inferring sleep...
SourceID proquest
pubmed
SourceType Aggregation Database
Index Database
Title Self-Supervised Representation Learning and Temporal-Spectral Feature Fusion for Bed Occupancy Detection
URI https://www.ncbi.nlm.nih.gov/pubmed/40093941
https://www.proquest.com/docview/3178292013
Volume 8
WOSCitedRecordID wos001318802900004&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV3JTsMwELWAcoAD-1I2GYmrRR072wmxVRygqmhBvVUTewJIVVqahu9nnKQqFyQkcsgtUeSMn9-Mn-cxdpGATqyCRKRAEKgtBAKk9gWmoCKDqUxK8fjrY9jpRINB3K0Lbnktq5xjYgnUdmxcjfyS1rnIOStJdTX5FM41yu2u1hYay6yhiMo4SVc4iBY1FsrWqxPTng61oFQgrM7NUhLgX6rAGdPr36llucS0N__7cVtsoyaX_LqKhm22hNkOW__RcnCXvfdwlIpeMXEYkaPlz6UUtj6BlPG63-obh8zyftW3aiScS70riXDHGIsp8nbhqmycGC-_oZdUvYoJp_kdzkpxV7bHXtr3_dsHUbstCFCt1kwoW_YK1HGIMrJ0gUHrxygxTFKiAUjEEnyarhiDMiATkMYYDyh_S0JDNG6frWTjDA8ZJ4AFlEFAVNIRNC9uQRqn2lojXVJsm-x8PoxDima3RQEZjot8uBjIJjuo_sVwUrXdGGpXfYm1PPrD08dszSP2UYnBTlgjpbmMp2zVfM0-8ulZGSZ073SfvgFe_cmR
linkProvider ProQuest
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Self-Supervised+Representation+Learning+and+Temporal-Spectral+Feature+Fusion+for+Bed+Occupancy+Detection&rft.jtitle=Proceedings+of+ACM+on+interactive%2C+mobile%2C+wearable+and+ubiquitous+technologies&rft.au=Song%2C+Yingjian&rft.au=Pitafi%2C+Zaid+Farooq&rft.au=Dou%2C+Fei&rft.au=Sun%2C+Jin&rft.date=2024-09-01&rft.issn=2474-9567&rft.eissn=2474-9567&rft.volume=8&rft.issue=3&rft_id=info:doi/10.1145%2F3678514&rft.externalDBID=NO_FULL_TEXT
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2474-9567&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2474-9567&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2474-9567&client=summon