001017949 001__ 1017949
001017949 005__ 20240403082757.0
001017949 0247_ $$2doi$$a10.1109/IGARSS52108.2023.10283284
001017949 0247_ $$2datacite_doi$$a10.34734/FZJ-2023-04454
001017949 0247_ $$2WOS$$aWOS:001098971605148
001017949 037__ $$aFZJ-2023-04454
001017949 1001_ $$0P:(DE-Juel1)178695$$aSedona, Rocco$$b0$$eCorresponding author$$ufzj
001017949 1112_ $$aIEEE International Geoscience and Remote Sensing Symposium (IGARSS)$$cPasadena$$d2023-07-16 - 2023-07-21$$wCA
001017949 245__ $$aEnhancing Training Set Through Multi-Temporal Attention Analysis in Transformers for Multi-Year Land Cover Mapping
001017949 260__ $$bIEEE$$c2023
001017949 300__ $$a5411-5414
001017949 3367_ $$2ORCID$$aCONFERENCE_PAPER
001017949 3367_ $$033$$2EndNote$$aConference Paper
001017949 3367_ $$2BibTeX$$aINPROCEEDINGS
001017949 3367_ $$2DRIVER$$aconferenceObject
001017949 3367_ $$2DataCite$$aOutput Types/Conference Paper
001017949 3367_ $$0PUB:(DE-HGF)8$$2PUB:(DE-HGF)$$aContribution to a conference proceedings$$bcontrib$$mcontrib$$s1704984848_2197
001017949 520__ $$aThe continuous stream of high spatial resolution satellite data offers the opportunity to regularly produce land cover (LC) maps. To this end, Transformer deep learning (DL) models have recently proven their effectiveness in accurately classifying long time series (TS) of satellite images. The continual generation of regularly updated LC maps can be used to analyze dynamic phenomena and extract multi-temporal information. However, several challenges need to be addressed. Our paper aims to study how the performance of a Transformer model changes when classifying TS of satellite images acquired in years later than those in the training set. In particular, the behavior of the attention in the Transformer model is analyzed to determine when the information provided by the initial training set needs to be updated to keep generating accurate LC products. Preliminary results show that: (i) the selection of the positional encoding strategy used in the Transformer has a significant impact on the classification accuracy obtained with multi-year TS, and (ii) the most affected classes are the seasonal ones.
001017949 536__ $$0G:(DE-HGF)POF4-5111$$a5111 - Domain-Specific Simulation & Data Life Cycle Labs (SDLs) and Research Groups (POF4-511)$$cPOF4-511$$fPOF IV$$x0
001017949 536__ $$0G:(EU-Grant)951733$$aRAISE - Research on AI- and Simulation-Based Engineering at Exascale (951733)$$c951733$$fH2020-INFRAEDI-2019-1$$x1
001017949 536__ $$0G:(DE-Juel-1)DEA02266$$aEUROCC-2 (DEA02266)$$cDEA02266$$x2
001017949 588__ $$aDataset connected to CrossRef Conference
001017949 7001_ $$0P:(DE-Juel1)187002$$aEbert, Jan$$b1$$ufzj
001017949 7001_ $$0P:(DE-HGF)0$$aParis, Claudia$$b2
001017949 7001_ $$0P:(DE-Juel1)132239$$aRiedel, Morris$$b3$$ufzj
001017949 7001_ $$0P:(DE-Juel1)171343$$aCavallaro, Gabriele$$b4$$ufzj
001017949 773__ $$a10.1109/IGARSS52108.2023.10283284
001017949 8564_ $$uhttps://juser.fz-juelich.de/record/1017949/files/Enhancing_Training_Set_through_Multi_temporal_Attention_Analysis_in_Transformers_for_multi_year_Land_Cover_Mapping.pdf$$yOpenAccess
001017949 8564_ $$uhttps://juser.fz-juelich.de/record/1017949/files/Enhancing_Training_Set_through_Multi_temporal_Attention_Analysis_in_Transformers_for_multi_year_Land_Cover_Mapping.gif?subformat=icon$$xicon$$yOpenAccess
001017949 8564_ $$uhttps://juser.fz-juelich.de/record/1017949/files/Enhancing_Training_Set_through_Multi_temporal_Attention_Analysis_in_Transformers_for_multi_year_Land_Cover_Mapping.jpg?subformat=icon-1440$$xicon-1440$$yOpenAccess
001017949 8564_ $$uhttps://juser.fz-juelich.de/record/1017949/files/Enhancing_Training_Set_through_Multi_temporal_Attention_Analysis_in_Transformers_for_multi_year_Land_Cover_Mapping.jpg?subformat=icon-180$$xicon-180$$yOpenAccess
001017949 8564_ $$uhttps://juser.fz-juelich.de/record/1017949/files/Enhancing_Training_Set_through_Multi_temporal_Attention_Analysis_in_Transformers_for_multi_year_Land_Cover_Mapping.jpg?subformat=icon-640$$xicon-640$$yOpenAccess
001017949 909CO $$ooai:juser.fz-juelich.de:1017949$$pdnbdelivery$$pec_fundedresources$$pVDB$$pdriver$$popen_access$$popenaire
001017949 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)178695$$aForschungszentrum Jülich$$b0$$kFZJ
001017949 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)187002$$aForschungszentrum Jülich$$b1$$kFZJ
001017949 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)132239$$aForschungszentrum Jülich$$b3$$kFZJ
001017949 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)171343$$aForschungszentrum Jülich$$b4$$kFZJ
001017949 9131_ $$0G:(DE-HGF)POF4-511$$1G:(DE-HGF)POF4-510$$2G:(DE-HGF)POF4-500$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$9G:(DE-HGF)POF4-5111$$aDE-HGF$$bKey Technologies$$lEngineering Digital Futures – Supercomputing, Data Management and Information Security for Knowledge and Action$$vEnabling Computational- & Data-Intensive Science and Engineering$$x0
001017949 9141_ $$y2023
001017949 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
001017949 9201_ $$0I:(DE-Juel1)JSC-20090406$$kJSC$$lJülich Supercomputing Center$$x0
001017949 980__ $$acontrib
001017949 980__ $$aVDB
001017949 980__ $$aUNRESTRICTED
001017949 980__ $$aI:(DE-Juel1)JSC-20090406
001017949 9801_ $$aFullTexts