Show simple item record

Files in this item

Thumbnail

Item metadata

dc.contributor.authorSchiemer, Martin
dc.contributor.authorFang, Lei
dc.contributor.authorDobson, Simon
dc.contributor.authorYe, Juan
dc.date.accessioned2023-07-28T10:30:04Z
dc.date.available2023-07-28T10:30:04Z
dc.date.issued2023-07-08
dc.identifier289930975
dc.identifierf819c045-7a58-4222-9218-9066604f1894
dc.identifier85164212115
dc.identifier.citationSchiemer , M , Fang , L , Dobson , S & Ye , J 2023 , ' Online continual learning for human activity recognition ' , Pervasive and Mobile Computing , vol. 93 , 101817 . https://doi.org/10.1016/j.pmcj.2023.101817en
dc.identifier.issn1574-1192
dc.identifier.otherRIS: urn:89243EA498334CF54A17DB1F99025AB5
dc.identifier.otherORCID: /0000-0002-2838-6836/work/138327307
dc.identifier.otherORCID: /0000-0001-9633-2103/work/138327339
dc.identifier.urihttps://hdl.handle.net/10023/28053
dc.descriptionFunding: This work is partly funded by Leverhulme Research Project Grant RPG-2021-355.en
dc.description.abstractSensor-based human activity recognition (HAR), with the ability to recognise human activities from wearable or embedded sensors, has been playing an important role in many applications including personal health monitoring, smart home, and manufacturing. The real-world, long-term deployment of these HAR systems drives a critical research question: how to evolve the HAR model automatically over time to accommodate changes in an environment or activity patterns. This paper presents an online continual learning (OCL) scenario for HAR, where sensor data arrives in a streaming manner which contains unlabelled samples from already learnt activities or new activities. We propose a technique, OCL-HAR, making a real-time prediction on the streaming sensor data while at the same time discovering and learning new activities. We have empirically evaluated OCL-HAR on four third-party, publicly available HAR datasets. Our results have shown that this OCL scenario is challenging to state-of-the-art continual learning techniques that have significantly underperformed. Our technique OCL-HAR has consistently outperformed them in all experiment setups, leading up to 0.17 and 0.23 improvements in micro and macro F1 scores.
dc.format.extent20
dc.format.extent1609408
dc.language.isoeng
dc.relation.ispartofPervasive and Mobile Computingen
dc.subjectHuman activity recognitionen
dc.subjectOnline continual learningen
dc.subjectDeep learningen
dc.subjectPervasive computingen
dc.subjectQA75 Electronic computers. Computer scienceen
dc.subject3rd-DASen
dc.subjectSDG 3 - Good Health and Well-beingen
dc.subjectACen
dc.subjectMCCen
dc.subject.lccQA75en
dc.titleOnline continual learning for human activity recognitionen
dc.typeJournal articleen
dc.contributor.sponsorThe Leverhulme Trusten
dc.contributor.institutionUniversity of St Andrews. School of Computer Scienceen
dc.contributor.institutionUniversity of St Andrews. Sir James Mackenzie Institute for Early Diagnosisen
dc.identifier.doi10.1016/j.pmcj.2023.101817
dc.description.statusPeer revieweden
dc.identifier.grantnumberRPG-2021-355en


This item appears in the following Collection(s)

Show simple item record