Publications
Detailed Information
Heterogeneous Double-Head Ensemble for Deep Metric Learning
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Ro, Youngmin | - |
dc.contributor.author | Choi, Jin Young | - |
dc.date.accessioned | 2023-12-11T00:35:15Z | - |
dc.date.available | 2023-12-11T00:35:15Z | - |
dc.date.created | 2020-08-12 | - |
dc.date.created | 2020-08-12 | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | IEEE Access, Vol.8, pp.118525-118533 | - |
dc.identifier.issn | 2169-3536 | - |
dc.identifier.uri | https://hdl.handle.net/10371/197660 | - |
dc.description.abstract | The structure of a multi-head ensemble has been employed by many algorithms in various applications including deep metric learning. However, their structures have been empirically designed in a simple way such as using the same head structure, which leads to a limited ensemble effect due to lack of head diversity. In this paper, for an elaborate design of the multi-head ensemble structure, we establish design concepts based on three structural factors: designing the feature layer for extracting the ensemble-favorable feature vector, designing the shared part for memory savings, and designing the diverse multi-heads for performance improvement. Through rigorous evaluation of variants on the basis of the design concepts, we propose a heterogeneous double-head ensemble structure that drastically increases ensemble gain along with memory savings. In verifying experiments on image retrieval datasets, the proposed ensemble structure outperforms the state-of-the-art algorithms by margins of over 5.3%, 6.1%, 5.9%, and 1.8% in CUB-200, Car-196, SOP, and Inshop, respectively. | - |
dc.language | 영어 | - |
dc.publisher | Institute of Electrical and Electronics Engineers Inc. | - |
dc.title | Heterogeneous Double-Head Ensemble for Deep Metric Learning | - |
dc.type | Article | - |
dc.identifier.doi | 10.1109/ACCESS.2020.3004579 | - |
dc.citation.journaltitle | IEEE Access | - |
dc.identifier.wosid | 000551850500001 | - |
dc.identifier.scopusid | 2-s2.0-85088129536 | - |
dc.citation.endpage | 118533 | - |
dc.citation.startpage | 118525 | - |
dc.citation.volume | 8 | - |
dc.description.isOpenAccess | Y | - |
dc.contributor.affiliatedAuthor | Choi, Jin Young | - |
dc.type.docType | Article | - |
dc.description.journalClass | 1 | - |
dc.subject.keywordAuthor | Ensemble learning | - |
dc.subject.keywordAuthor | multi-head structure | - |
dc.subject.keywordAuthor | deep metric learning | - |
dc.subject.keywordAuthor | deep architecture design | - |
dc.subject.keywordAuthor | image retrieval | - |
- Appears in Collections:
- Files in This Item:
- There are no files associated with this item.
Item View & Download Count
Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.