Detailed Information

Cited 0 time in webofscience Cited 0 time in scopus
Metadata Downloads

ACP++: Action Co-occurrence Priors for Human-Object Interaction Detection

Full metadata record
DC Field Value Language
dc.contributor.authorKim, Dong Jin-
dc.contributor.authorSun, Xiao-
dc.contributor.authorChoi, Jinsoo-
dc.contributor.authorLin, Stephen-
dc.contributor.authorKweon, In So-
dc.date.accessioned2023-08-16T08:04:01Z-
dc.date.available2023-08-16T08:04:01Z-
dc.date.created2023-07-21-
dc.date.issued2021-09-
dc.identifier.issn1057-7149-
dc.identifier.urihttps://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/189248-
dc.description.abstractA common problem in the task of human-object interaction (HOI) detection is that numerous HOI classes have only a small number of labeled examples, resulting in training sets with a long-tailed distribution. The lack of positive labels can lead to low classification accuracy for these classes. Towards addressing this issue, we observe that there exist natural correlations and anti-correlations among human-object interactions. In this paper, we model the correlations as action co-occurrence matrices and present techniques to learn these priors and leverage them for more effective training, especially on rare classes. The efficacy of our approach is demonstrated experimentally, where the performance of our approach consistently improves over the state-of-the-art methods on both of the two leading HOI detection benchmark datasets, HICO-Det and V-COCO.-
dc.language영어-
dc.language.isoen-
dc.publisherIEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC-
dc.titleACP++: Action Co-occurrence Priors for Human-Object Interaction Detection-
dc.typeArticle-
dc.contributor.affiliatedAuthorKim, Dong Jin-
dc.identifier.doi10.1109/TIP.2021.3113563-
dc.identifier.scopusid2-s2.0-85115758158-
dc.identifier.wosid000716696700003-
dc.identifier.bibliographicCitationIEEE TRANSACTIONS ON IMAGE PROCESSING, v.30, pp.9150 - 9163-
dc.relation.isPartOfIEEE TRANSACTIONS ON IMAGE PROCESSING-
dc.citation.titleIEEE TRANSACTIONS ON IMAGE PROCESSING-
dc.citation.volume30-
dc.citation.startPage9150-
dc.citation.endPage9163-
dc.type.rimsART-
dc.type.docType정기학술지(Article(Perspective Article포함))-
dc.description.journalClass1-
dc.description.isOpenAccessY-
dc.description.journalRegisteredClassscie-
dc.description.journalRegisteredClassscopus-
dc.relation.journalResearchAreaComputer Science-
dc.relation.journalResearchAreaEngineering-
dc.relation.journalWebOfScienceCategoryComputer Science, Artificial Intelligence-
dc.relation.journalWebOfScienceCategoryEngineering, Electrical & Electronic-
dc.subject.keywordPlusCLASSIFICATION-
dc.subject.keywordAuthorVisualization-
dc.subject.keywordAuthorTraining-
dc.subject.keywordAuthorTask analysis-
dc.subject.keywordAuthorBicycles-
dc.subject.keywordAuthorSemantics-
dc.subject.keywordAuthorContext modeling-
dc.subject.keywordAuthorBenchmark testing-
dc.subject.keywordAuthorHuman-object interaction-
dc.subject.keywordAuthorVisual relationship-
dc.subject.keywordAuthorCo-occurrence-
dc.subject.keywordAuthorLabel hierarchy-
dc.subject.keywordAuthorKnowledge distillation-
dc.identifier.urlhttps://ieeexplore.ieee.org/document/9547056-
Files in This Item
Go to Link
Appears in
Collections
ETC > 1. Journal Articles

qrcode

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.

Related Researcher

Researcher Kim, Dong Jin photo

Kim, Dong Jin
COLLEGE OF ENGINEERING (DEPARTMENT OF INTELLIGENCE COMPUTING)
Read more

Altmetrics

Total Views & Downloads

BROWSE