Integrating Multiple Receptive Fields Through Grouped Active Convolution

Cited 1 time in webofscience Cited 0 time in scopus
  • Hit : 267
  • Download : 9
DC FieldValueLanguage
dc.contributor.authorJeon, Yunhoko
dc.contributor.authorKim, Junmoko
dc.date.accessioned2021-10-18T06:30:07Z-
dc.date.available2021-10-18T06:30:07Z-
dc.date.created2021-10-18-
dc.date.created2021-10-18-
dc.date.created2021-10-18-
dc.date.issued2021-11-
dc.identifier.citationIEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, v.43, no.11, pp.3892 - 3903-
dc.identifier.issn0162-8828-
dc.identifier.urihttp://hdl.handle.net/10203/288223-
dc.description.abstractConvolutional networks have achieved great success in various vision tasks. This is mainly due to a considerable amount of research on network structure. In this study, instead of focusing on architectures, we focused on the convolution unit itself. The existing convolution unit has a fixed shape and is limited to observing restricted receptive fields. In earlier work, we proposed the active convolution unit (ACU), which can freely define its shape and learn by itself. In this paper, we provide a detailed analysis of the previously proposed unit and show that it is an efficient representation of a sparse weight convolution. Furthermore, we extend an ACU to a grouped ACU, which can observe multiple receptive fields in one layer. We found that the performance of a naive grouped convolution is degraded by increasing the number of groups; however, the proposed unit retains the accuracy even though the number of parameters decreases. Based on this result, we suggest a depthwise ACU (DACU), and various experiments have shown that our unit is efficient and can replace the existing convolutions.-
dc.languageEnglish-
dc.publisherIEEE COMPUTER SOC-
dc.titleIntegrating Multiple Receptive Fields Through Grouped Active Convolution-
dc.typeArticle-
dc.identifier.wosid000702649700015-
dc.identifier.scopusid2-s2.0-85116534039-
dc.type.rimsART-
dc.citation.volume43-
dc.citation.issue11-
dc.citation.beginningpage3892-
dc.citation.endingpage3903-
dc.citation.publicationnameIEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE-
dc.identifier.doi10.1109/TPAMI.2020.2995864-
dc.embargo.liftdate9999-12-31-
dc.embargo.terms9999-12-31-
dc.contributor.localauthorKim, Junmo-
dc.description.isOpenAccessN-
dc.type.journalArticleArticle-
dc.subject.keywordAuthorConvolution-
dc.subject.keywordAuthorShape-
dc.subject.keywordAuthorTask analysis-
dc.subject.keywordAuthorComputer architecture-
dc.subject.keywordAuthorSemantics-
dc.subject.keywordAuthorNetwork architecture-
dc.subject.keywordAuthorBackpropagation-
dc.subject.keywordAuthorConvolutional neural network (CNN)-
dc.subject.keywordAuthormultiple receptive fields-
dc.subject.keywordAuthordepthwise convolution-
dc.subject.keywordAuthordeep learning-
Appears in Collection
EE-Journal Papers(저널논문)
Files in This Item
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 1 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0