dc.identifier.uri | http://dx.doi.org/10.15488/1179 | |
dc.identifier.uri | http://www.repo.uni-hannover.de/handle/123456789/1203 | |
dc.contributor.author | Paul, Andreas | |
dc.contributor.author | Rottensteiner, Franz | |
dc.contributor.author | Heipke, Christian | |
dc.contributor.editor | Halounova, L. | |
dc.contributor.editor | Schindler, K. | |
dc.contributor.editor | Limpouch, A. | |
dc.contributor.editor | Pajdla, T. | |
dc.contributor.editor | Šafář, V. | |
dc.contributor.editor | Mayer, H. | |
dc.contributor.editor | Oude Elberink, S. | |
dc.contributor.editor | Mallet, C. | |
dc.contributor.editor | Rottensteiner, F. | |
dc.contributor.editor | Brédif, M. | |
dc.contributor.editor | Skaloud, J. | |
dc.contributor.editor | Stilla, U. | |
dc.date.accessioned | 2017-03-02T12:48:02Z | |
dc.date.available | 2017-03-02T12:48:02Z | |
dc.date.issued | 2016 | |
dc.identifier.citation | Paul, A.; Rottensteiner, F.; Heipke, C.: Iterative re-weighted instance transfer for domain adaptation. In: XXIII ISPRS Congress, Commission III 3 (2016), Nr. 3, S. 339-346. DOI: https://doi.org/10.5194/isprsannals-III-3-339-2016 | |
dc.description.abstract | Domain adaptation techniques in transfer learning try to reduce the amount of training data required for classification by adapting a classifier trained on samples from a source domain to a new data set (target domain) where the features may have different distributions. In this paper, we propose a new technique for domain adaptation based on logistic regression. Starting with a classifier trained on training data from the source domain, we iteratively include target domain samples for which class labels have been obtained from the current state of the classifier, while at the same time removing source domain samples. In each iteration the classifier is re-trained, so that the decision boundaries are slowly transferred to the distribution of the target features. To make the transfer procedure more robust we introduce weights as a function of distance from the decision boundary and a new way of regularisation. Our methodology is evaluated using a benchmark data set consisting of aerial images and digital surface models. The experimental results show that in the majority of cases our domain adaptation approach can lead to an improvement of the classification accuracy without additional training data, but also indicate remaining problems if the difference in the feature distributions becomes too large. | eng |
dc.description.sponsorship | DFG/HE 1822/30-1 | |
dc.language.iso | eng | |
dc.publisher | Göttingen : Copernicus GmbH | |
dc.relation.ispartof | XXIIIrd ISPRS congress 2016 : Prague, Czech Republic, 12th-19th July 2016 | |
dc.relation.ispartofseries | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences ; III-3 | |
dc.rights | CC BY 3.0 Unported | |
dc.rights.uri | https://creativecommons.org/licenses/by/3.0/ | |
dc.subject | transfer learning | eng |
dc.subject | domain adaptation | eng |
dc.subject | logistic regression | eng |
dc.subject | machine learning | eng |
dc.subject | knowledge transfer | eng |
dc.subject | remote sensing | eng |
dc.subject | remote-sensing images | eng |
dc.subject | validation | eng |
dc.subject | classifier | eng |
dc.subject.classification | Konferenzschrift | ger |
dc.subject.ddc | 550 | Geowissenschaften | ger |
dc.title | Iterative re-weighted instance transfer for domain adaptation | |
dc.type | Article | |
dc.type | Text | |
dc.relation.essn | 2194-9050 | |
dc.relation.issn | 2194-9034 | |
dc.relation.doi | https://doi.org/10.5194/isprsannals-III-3-339-2016 | |
dc.bibliographicCitation.issue | 3 | |
dc.bibliographicCitation.volume | III-3 | |
dc.bibliographicCitation.firstPage | 339 | |
dc.bibliographicCitation.lastPage | 346 | |
dc.description.version | publishedVersion | |
tib.accessRights | frei zug�nglich |
Die folgenden Lizenzbestimmungen sind mit dieser Ressource verbunden: