DC Field | Value | Language |
---|---|---|
dc.contributor.author | Choi, Seung-Min | ko |
dc.contributor.author | Lee, Seung-Ik | ko |
dc.contributor.author | Lee, Jae-Yeong | ko |
dc.contributor.author | Kweon, In So | ko |
dc.date.accessioned | 2023-06-21T07:00:12Z | - |
dc.date.available | 2023-06-21T07:00:12Z | - |
dc.date.created | 2023-06-21 | - |
dc.date.issued | 2023-09 | - |
dc.identifier.citation | PATTERN RECOGNITION, v.141 | - |
dc.identifier.issn | 0031-3203 | - |
dc.identifier.uri | http://hdl.handle.net/10203/307421 | - |
dc.description.abstract | Thanks to Earth-level Street View images from Google Maps, a visual image geo-localization can estimate the coarse location of a query image with a visual place recognition process. However, this can get very challenging when non-static objects change with time, severely degrading image retrieval accuracy. We address the problem of city-scale visual place recognition in complex urban environments crowded with non-static clutters. To this end, we first analyze what clutters degrade similarity matching between the query and database images. Second, we design a self-supervised trainable de-attention module that pre-vents the network from focusing on non-static objects in an input image. In addition, we propose a novel triplet marginal loss called sharpened triplet marginal loss to make feature descriptors more discriminative. Lastly, due to the lack of geo-tagged public datasets with a high density of non-static objects, we propose a clutter augmentation method to evaluate our approach. The experimental results show that our model has notably improved over the existing attention methods in geo-localization tasks on the public bench-mark datasets and on their augmented versions with high population and traffic. Our code is available at https://github.com/ccsmm78/deattention _ with _ stml _ for _ vpr .(c) 2023 The Author(s). Published by Elsevier Ltd.This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ) | - |
dc.language | English | - |
dc.publisher | ELSEVIER SCI LTD | - |
dc.title | Semantic-guided de-attention with sharpened triplet marginal loss for visual place recognition | - |
dc.type | Article | - |
dc.identifier.wosid | 000999041000001 | - |
dc.identifier.scopusid | 2-s2.0-85156187545 | - |
dc.type.rims | ART | - |
dc.citation.volume | 141 | - |
dc.citation.publicationname | PATTERN RECOGNITION | - |
dc.identifier.doi | 10.1016/j.patcog.2023.109645 | - |
dc.contributor.localauthor | Kweon, In So | - |
dc.contributor.nonIdAuthor | Lee, Seung-Ik | - |
dc.contributor.nonIdAuthor | Lee, Jae-Yeong | - |
dc.description.isOpenAccess | N | - |
dc.type.journalArticle | Article | - |
dc.subject.keywordAuthor | Visual place recognition | - |
dc.subject.keywordAuthor | Image retrieval | - |
dc.subject.keywordAuthor | Triplet marginal loss | - |
dc.subject.keywordAuthor | Attention | - |
dc.subject.keywordAuthor | De-attention | - |
dc.subject.keywordAuthor | Semantic guidance | - |
dc.subject.keywordAuthor | Semantic segmentation | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.