000851270 001__ 851270
000851270 005__ 20210129234826.0
000851270 0247_ $$2Handle$$a2128/19649
000851270 037__ $$aFZJ-2018-04965
000851270 1001_ $$0P:(DE-HGF)0$$aLange, Julius$$b0
000851270 1112_ $$aIEEE International Geoscience and Remote Sensing Symposium (IGARSS)$$cValencia$$d2018-07-22 - 2018-07-27$$gIGARSS$$wSpain
000851270 245__ $$aThe Influence of Sampling Methods on Pixel-Wise Hyperspectral Image Classification with 3D Convolutional Neural Networks
000851270 260__ $$c2018
000851270 3367_ $$033$$2EndNote$$aConference Paper
000851270 3367_ $$2DataCite$$aOther
000851270 3367_ $$2BibTeX$$aINPROCEEDINGS
000851270 3367_ $$2DRIVER$$aconferenceObject
000851270 3367_ $$2ORCID$$aLECTURE_SPEECH
000851270 3367_ $$0PUB:(DE-HGF)6$$2PUB:(DE-HGF)$$aConference Presentation$$bconf$$mconf$$s1536322046_27185$$xInvited
000851270 520__ $$aSupervised image classification is one of the essential techniques for generating semantic maps from remotely sensed images. The lack of labeled ground truth datasets, due to the inherent time effort and cost involved in collecting training samples, has led to the practice of training and validating new classifiers within a single image. In line with that, the dominant approach for the division of the available ground truth into disjoint training and test sets is random sampling. This paper discusses the problems that arise when this strategy is adopted in conjunction with spectral-spatial and pixel-wise classifiers such as 3D Convolutional Neural Networks (3D CNN). It is shown that a random sampling scheme leads to a violation of the independence assumption and to the illusion that global knowledge is extracted from the training set.To tackle this issue, two improved sampling strategies based on the Density-Based Clustering Algorithm (DBSCAN) are proposed. They minimize the violation of the train and test samples independence assumption and thus ensure an honest estimation of the generalization capabilities of the classifier.
000851270 536__ $$0G:(DE-HGF)POF3-512$$a512 - Data-Intensive Science and Federated Computing (POF3-512)$$cPOF3-512$$fPOF III$$x0
000851270 536__ $$0G:(EU-Grant)754304$$aDEEP-EST - DEEP - Extreme Scale Technologies (754304)$$c754304$$fH2020-FETHPC-2016$$x1
000851270 536__ $$0G:(EU-Grant)763558$$aSIMDAS - Upgrade of CaSToRC into a Center of Excellence in Simulation and Data Science (763558)$$c763558$$fH2020-WIDESPREAD-04-2017-TeamingPhase1$$x2
000851270 7001_ $$0P:(DE-Juel1)171343$$aCavallaro, Gabriele$$b1$$eCorresponding author
000851270 7001_ $$0P:(DE-Juel1)162390$$aGötz, Markus$$b2
000851270 7001_ $$0P:(DE-HGF)0$$aErlingsson, Ernir$$b3
000851270 7001_ $$0P:(DE-Juel1)132239$$aRiedel, Morris$$b4
000851270 8564_ $$uhttps://juser.fz-juelich.de/record/851270/files/IGARSS_2018_Cavallaro.pdf$$yOpenAccess
000851270 8564_ $$uhttps://juser.fz-juelich.de/record/851270/files/IGARSS_2018_Cavallaro.gif?subformat=icon$$xicon$$yOpenAccess
000851270 8564_ $$uhttps://juser.fz-juelich.de/record/851270/files/IGARSS_2018_Cavallaro.jpg?subformat=icon-1440$$xicon-1440$$yOpenAccess
000851270 8564_ $$uhttps://juser.fz-juelich.de/record/851270/files/IGARSS_2018_Cavallaro.jpg?subformat=icon-180$$xicon-180$$yOpenAccess
000851270 8564_ $$uhttps://juser.fz-juelich.de/record/851270/files/IGARSS_2018_Cavallaro.jpg?subformat=icon-640$$xicon-640$$yOpenAccess
000851270 8564_ $$uhttps://juser.fz-juelich.de/record/851270/files/IGARSS_2018_Cavallaro.pdf?subformat=pdfa$$xpdfa$$yOpenAccess
000851270 909CO $$ooai:juser.fz-juelich.de:851270$$pec_fundedresources$$pdriver$$pVDB$$popen_access$$popenaire
000851270 9101_ $$0I:(DE-HGF)0$$6P:(DE-HGF)0$$a Humboldt University Berlin$$b0
000851270 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)171343$$aForschungszentrum Jülich$$b1$$kFZJ
000851270 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)162390$$aForschungszentrum Jülich$$b2$$kFZJ
000851270 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)132239$$aForschungszentrum Jülich$$b4$$kFZJ
000851270 9131_ $$0G:(DE-HGF)POF3-512$$1G:(DE-HGF)POF3-510$$2G:(DE-HGF)POF3-500$$3G:(DE-HGF)POF3$$4G:(DE-HGF)POF$$aDE-HGF$$bKey Technologies$$lSupercomputing & Big Data$$vData-Intensive Science and Federated Computing$$x0
000851270 9141_ $$y2018
000851270 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
000851270 9201_ $$0I:(DE-Juel1)JSC-20090406$$kJSC$$lJülich Supercomputing Center$$x0
000851270 980__ $$aconf
000851270 980__ $$aVDB
000851270 980__ $$aUNRESTRICTED
000851270 980__ $$aI:(DE-Juel1)JSC-20090406
000851270 9801_ $$aFullTexts