Deep domain adaptation by weighted entropy minimization for the classification of aerial images

Show simple item record

dc.identifier.uri http://dx.doi.org/10.15488/10881
dc.identifier.uri https://www.repo.uni-hannover.de/handle/123456789/10963
dc.contributor.author Wittich, Dennis
dc.contributor.editor Paparoditis, N.
dc.contributor.editor Mallet, C.
dc.contributor.editor Lafarge, F.
dc.contributor.editor Remondino, F.
dc.contributor.editor Toschi, I.
dc.contributor.editor Fuse, T.
dc.date.accessioned 2021-05-04T12:14:03Z
dc.date.available 2021-05-04T12:14:03Z
dc.date.issued 2020
dc.identifier.citation Wittich, D.: Deep domain adaptation by weighted entropy minimization for the classification of aerial images. In: Paparoditis, N. et.al. (Eds.): XXIV ISPRS Congress, Commission II : edition 2020. Katlenburg-Lindau : Copernicus Publications, 2020 (ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences ; 5,2), S. 591-598. DOI: https://doi.org/10.5194/isprs-annals-V-2-2020-591-2020
dc.description.abstract Fully convolutional neural networks (FCN) are successfully used for the automated pixel-wise classification of aerial images and possibly additional data. However, they require many labelled training samples to perform well. One approach addressing this issue is semi-supervised domain adaptation (SSDA). Here, labelled training samples from a source domain and unlabelled samples from a target domain are used jointly to obtain a target domain classifier, without requiring any labelled samples from the target domain. In this paper, a two-step approach for SSDA is proposed. The first step corresponds to a supervised training on the source domain, making use of strong data augmentation to increase the initial performance on the target domain. Secondly, the model is adapted by entropy minimization using a novel weighting strategy. The approach is evaluated on the basis of five domains, corresponding to five cities. Several training variants and adaptation scenarios are tested, indicating that proper data augmentation can already improve the initial target domain performance significantly resulting in an average overall accuracy of 77.5%. The weighted entropy minimization improves the overall accuracy on the target domains in 19 out of 20 scenarios on average by 1.8%. In all experiments a novel FCN architecture is used that yields results comparable to those of the best-performing models on the ISPRS labelling challenge while having an order of magnitude fewer parameters than commonly used FCNs. © 2020 Copernicus GmbH. All rights reserved. eng
dc.language.iso eng
dc.publisher Katlenburg-Lindau : Copernicus Publications
dc.relation.ispartof XXIV ISPRS Congress, Commission II : edition 2020
dc.relation.ispartofseries ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences ; 5,2
dc.rights CC BY 4.0 Unported
dc.rights.uri https://creativecommons.org/licenses/by/4.0/
dc.subject domain adaptation eng
dc.subject aerial images eng
dc.subject classification eng
dc.subject fully convolutional networks eng
dc.subject entropy minimization eng
dc.subject.classification Konferenzschrift ger
dc.subject.ddc 550 | Geowissenschaften ger
dc.title Deep domain adaptation by weighted entropy minimization for the classification of aerial images
dc.type bookPart
dc.type Text
dc.relation.essn 2194-9050
dc.relation.issn 2194-9042
dc.relation.doi https://doi.org/10.5194/isprs-annals-V-2-2020-591-2020
dc.bibliographicCitation.issue 2
dc.bibliographicCitation.volume 5
dc.bibliographicCitation.firstPage 591
dc.bibliographicCitation.lastPage 598
dc.description.version publishedVersion
tib.accessRights frei zug�nglich


Files in this item

This item appears in the following Collection(s):

Show simple item record

 

Search the repository


Browse

My Account

Usage Statistics