Class-Conditioned Variational Autoencoder with Evolutionary Optimization for the Virtual Data Generation Challenge

Human activity recognition (HAR) plays a crucial role in optimizing packing operations in industrial settings. However, HAR performance is often constrained by the limited availability of labeled data. To address this issue, we propose a novel data augmentation framework using conditional variationa...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:International Journal of Activity and Behavior Computing Ročník 2025; číslo 2; s. 1 - 14
Hlavný autor: Tachioka, Yuuki
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Care XDX Center, Kyushu Institute of Technology 22.05.2025
ISSN:2759-2871
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract Human activity recognition (HAR) plays a crucial role in optimizing packing operations in industrial settings. However, HAR performance is often constrained by the limited availability of labeled data. To address this issue, we propose a novel data augmentation framework using conditional variational autoencoders (CVAE) to generate high-quality synthetic data. Our approach ensures class consistency while increasing data diversity by conditioning the generation process on activity labels. Additionally, we optimize hyperparameters for data generation using evolutionary computation, further improving recognition accuracy. The proposed method is validated on the OpenPack dataset, demonstrating its effectiveness in enhancing HAR performance without modifying the recognition model itself. Our key contributions include the introduction of a robust data augmentation pipeline, the application of CVAE for HAR, and the use of evolutionary computation to optimize data generation. Our model trained with data augmentation achieved an F1 score of 53.4%, while the recognition model trained without data augmentation achieved an F1 score of 48.1%.
AbstractList Human activity recognition (HAR) plays a crucial role in optimizing packing operations in industrial settings. However, HAR performance is often constrained by the limited availability of labeled data. To address this issue, we propose a novel data augmentation framework using conditional variational autoencoders (CVAE) to generate high-quality synthetic data. Our approach ensures class consistency while increasing data diversity by conditioning the generation process on activity labels. Additionally, we optimize hyperparameters for data generation using evolutionary computation, further improving recognition accuracy. The proposed method is validated on the OpenPack dataset, demonstrating its effectiveness in enhancing HAR performance without modifying the recognition model itself. Our key contributions include the introduction of a robust data augmentation pipeline, the application of CVAE for HAR, and the use of evolutionary computation to optimize data generation. Our model trained with data augmentation achieved an F1 score of 53.4%, while the recognition model trained without data augmentation achieved an F1 score of 48.1%.
Author Tachioka, Yuuki
Author_xml – sequence: 1
  fullname: Tachioka, Yuuki
  organization: Denso IT Laboratory
BookMark eNo9kM1OwzAQhC0EEqX0whP4BVL87_RYhVKQKvUCvUabxG5cuU7luCB4ekKDOM1I3-5oNHfoOnTBIPRAyVwRQeijO0BVzymRV2jCtFxkLNf0Fs363lVESr4ghJEJioWHvs-KLjQuuSGjwTuIDn49eLw8p86EumtMxJ8utXj10fnzBcYvvD0ld3Tfl2Nsu4hTa_DOxXQeXp8gAV6bYOLIixa8N2Fv7tGNBd-b2Z9O0fvz6q14yTbb9Wux3GQHylTILLWWWy2UYrVQuSW6AU2EWOS2roziQlDDLaHCcm6potJaUmldKaCQV5LwKSrG3EOfYG_KU3THoXUJMbnam_KyUMkIkyUbZVjrn9YtxNIE_gOWJ2q-
ContentType Journal Article
Copyright 2025 Author
Copyright_xml – notice: 2025 Author
DOI 10.60401/ijabc.105
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
EISSN 2759-2871
EndPage 14
ExternalDocumentID article_ijabc_2025_2_2025_105_article_char_en
GroupedDBID JSF
JSH
M~E
RJT
RZJ
ID FETCH-LOGICAL-j126n-f1ff3f74662c468f07da704498fcbe63441e3f014f33f1615ff0b77b6a1a8b503
IngestDate Wed Sep 03 06:30:28 EDT 2025
IsDoiOpenAccess true
IsOpenAccess true
IsPeerReviewed false
IsScholarly false
Issue 2
Language English
License https://creativecommons.org/licenses/by/4.0/deed.ja
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-j126n-f1ff3f74662c468f07da704498fcbe63441e3f014f33f1615ff0b77b6a1a8b503
OpenAccessLink https://www.jstage.jst.go.jp/article/ijabc/2025/2/2025_105/_article/-char/en
PageCount 14
ParticipantIDs jstage_primary_article_ijabc_2025_2_2025_105_article_char_en
PublicationCentury 2000
PublicationDate 2025/05/22
PublicationDateYYYYMMDD 2025-05-22
PublicationDate_xml – month: 05
  year: 2025
  text: 2025/05/22
  day: 22
PublicationDecade 2020
PublicationTitle International Journal of Activity and Behavior Computing
PublicationTitleAlternate IJABC
PublicationYear 2025
Publisher Care XDX Center, Kyushu Institute of Technology
Publisher_xml – name: Care XDX Center, Kyushu Institute of Technology
SSID ssib055390020
Score 1.9092373
Snippet Human activity recognition (HAR) plays a crucial role in optimizing packing operations in industrial settings. However, HAR performance is often constrained by...
SourceID jstage
SourceType Publisher
StartPage 1
Title Class-Conditioned Variational Autoencoder with Evolutionary Optimization for the Virtual Data Generation Challenge
URI https://www.jstage.jst.go.jp/article/ijabc/2025/2/2025_105/_article/-char/en
Volume 2025
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
ispartofPNX International Journal of Activity and Behavior Computing, 2025/05/22, Vol.2025(2), pp.1-14
journalDatabaseRights – providerCode: PRVHPJ
  databaseName: ROAD: Directory of Open Access Scholarly Resources
  databaseCode: M~E
  dateStart: 20240101
  customDbUrl:
  isFulltext: true
  eissn: 2759-2871
  dateEnd: 99991231
  titleUrlDefault: https://road.issn.org
  omitProxy: false
  ssIdentifier: ssib055390020
  providerName: ISSN International Centre
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV3Nb9MwFLfK4MAFgQDxLR-4VRGJ49iJxGXqirgwOIypnCLHtbV0LJ2yptou-5f4F3nPcVKzXZgQl6RK0r4075f35fdByHtjZc5NxqJEcSzJUTZSRa4ilVe2AJUvtVRu2IQ8PMwXi-LbZPJrqIXZ_pRNk19eFuf_ldVwDJiNpbN3YPf4o3AAPgPTYQtsh-1fMd6NuYxma1yLxjZEy-kx-MNDzG-_26yxdyW2kHAx2PnW3w2mz30FAXLmKzPHBMTjunVFJgdqo3yband-NsxhCQ3c2xHGwdzVfk4FRup9V8Z22g-VGNSniyDok3p96kzaH113WodhCZbhijrbObFYPTVdHCxckNqnh1x1Fyfdn0kQN9YPjBN8TGZFhJ5cKKWRRoBHFgjdJNDefUXqTb0gQFShYqhXqtI42nin_cacRM-w0l1TIrmS9Tu4vhzOYi0coOweuY93iTmEX67ngwDLsrRA-7tvg-uIfhhJgomzAoN_SBZ09svRY_LIc4Lu9xSekIlpnpL2FlhoABYagIUiWGgIFhqChQJYKICFerBQBAvdgYWOYHlGvn-aH80-R34KR7RKmGgim1ibWsmFYJqL3MZyqWTMeZFbXRmRgj1tUguetk1Ti_6DtXElZSVUAi98FqfPyV4Df-AFoYnQSwPPyKbgxjMQBVwwrrPYcLvEwNpL8rF_ROV532qlvBNTXv3b11-ThzsgvyF7m7Yzb8kDvd3UF-07x-jffWiV0g
linkProvider ISSN International Centre
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=Class-Conditioned+Variational+Autoencoder+with+Evolutionary+Optimization+for+the+Virtual+Data+Generation+Challenge&rft.jtitle=International+Journal+of+Activity+and+Behavior+Computing&rft.au=Tachioka%2C+Yuuki&rft.date=2025-05-22&rft.pub=Care+XDX+Center%2C+Kyushu+Institute+of+Technology&rft.eissn=2759-2871&rft.volume=2025&rft.issue=2&rft.spage=1&rft.epage=14&rft_id=info:doi/10.60401%2Fijabc.105&rft.externalDocID=article_ijabc_2025_2_2025_105_article_char_en