Please use this identifier to cite or link to this item: http://hdl.handle.net/10662/20346
Full metadata record
DC FieldValueLanguage
dc.contributor.authorPaoletti Ávila, Mercedes Eugenia-
dc.contributor.authorHaut Hurtado, Juan Mario-
dc.contributor.authorFernández Beltrán, Rubén-
dc.contributor.authorPlaza Miguel, Javier-
dc.contributor.authorPlaza, Antonio-
dc.contributor.authorPla, Filiberto-
dc.date.accessioned2024-02-07T13:04:34Z-
dc.date.available2024-02-07T13:04:34Z-
dc.date.issued2019-
dc.identifier.issn0196-2892-
dc.identifier.urihttp://hdl.handle.net/10662/20346-
dc.description.abstractConvolutional neural networks (CNNs) exhibit good performance in image processing tasks, pointing themselves as the current state-of-the-art of deep learning methods. However, the intrinsic complexity of remotely sensed hyperspectral images (HSIs) still limits the performance of many CNN models. The high dimensionality of HSI data, together with the underlying redundancy and noise, often make standard CNN approaches unable to generalize discriminative spectral-spatial features. Moreover, deeper CNN architectures also find challenges when additional layers are added, which hampers the network convergence and produces low classification accuracies. In order to mitigate these issues, this paper presents a new deep CNN architecture specially designed for HSI data. Our new model pursues to improve the spectral-spatial features uncovered by the convolutional filters of the network. Specifically, the proposed residual-based approach gradually increases the feature map dimension at all convolutional layers, grouped in pyramidal bottleneck residual blocks, in order to involve more locations as the network depth increases while balancing the workload among all units, preserving the time complexity per layer. It can be seen as a pyramid, where the deeper the blocks, the more feature maps can be extracted. Therefore, the diversity of high-level spectralspatial attributes can be gradually increased across layers to enhance the performance of the proposed network with HIS data. Our experiments, conducted using four well-known HIS datasets and ten different classification techniques, reveal that our newly developed HSI pyramidal residual model is able to provide competitive advantages (in terms of both classification accuracy and computational time) over state-of-the-art HSI classification methods.en_Us
dc.description.sponsorshipThis work has been supported by Ministerio de Education (Resolución de 26 de diciembre de 2014 y de 19 de noviembre de 2015, de la Secretaría de Estado de Educación, Formación Profesional y Universidades, por la que se convocan ayudas para la formación de profesorado universitario, de los subprogramas de Formación y de Movilidad incluidos en el Programa Estatal de Promoción del Talento y su Empleabilidad, en el marco del Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016. This work has also been supported by Junta de Extremadura (decreto 297/2014, ayudas para la realización de actividades de investigación y desarrollo tecnológico, de divulgación y de transferencia de conocimiento por los Grupos de Investigación de Extremadura, Ref. GR15005). This work has been additionally supported by the Generalitat Valenciana through the contract APOSTD/2017/007 and by the Spanish Ministry of Economy under the project ESP2016-79503-C2-2-P. (Corresponding author: M.E. Paoletti.).en_Us
dc.format.extent16 p.es_ES
dc.format.mimetypeapplication/pdfen_US
dc.language.isoenges_ES
dc.publisherIEEE-
dc.rightsAtribución 4.0 Internacional-
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/-
dc.subjectImágenes hiperespectraleses_ES
dc.subjectRedes neuronales convolucionaleses_ES
dc.subjectRedes residualeses_ES
dc.subjectHyperspectral imaging (HSI)en_Us
dc.subjectConvolutional neural networks (CNNs)en_US
dc.subjectResidual networks (ResNets)en_Us
dc.titleDeep pyramidal residual networks for spectral–spatial hyperspectral image classificationes_ES
dc.typearticlees_ES
dc.description.versionpeerReviewedes_ES
europeana.typeTEXTen_US
dc.rights.accessRightsopenAccesses_ES
dc.subject.unesco3304 Tecnología de Los Ordenadores-
europeana.dataProviderUniversidad de Extremadura. Españaes_ES
dc.identifier.bibliographicCitationM. E. Paoletti, J. M. Haut, R. Fernandez-Beltran, J. Plaza, A. J. Plaza and F. Pla, "Deep Pyramidal Residual Networks for Spectral–Spatial Hyperspectral Image Classification," in IEEE Transactions on Geoscience and Remote Sensing, vol. 57, no. 2, pp. 740-754, Feb. 2019, doi: 10.1109/TGRS.2018.2860125-
dc.type.versionpublishedVersion-
dc.contributor.affiliationUniversidad de Extremadura. Departamento de Tecnología de los Computadores y de las Comunicacioneses_ES
dc.contributor.affiliationUniversidad Jaume I-
dc.relation.publisherversionhttps://ieeexplore.ieee.org/document/8445697-
dc.identifier.doi10.1109/TGRS.2018.2860125-
dc.identifier.publicationtitleIEEE Transactions on Geoscience and Remote Sensinges_ES
dc.identifier.publicationissue2-
dc.identifier.publicationfirstpage740es_ES
dc.identifier.publicationlastpage754es_ES
dc.identifier.publicationvolume57es_ES
dc.identifier.e-issn1558-0644-
dc.identifier.orcid0000-0003-1030-3729es_ES
dc.identifier.orcid0000-0001-6701-961X-
dc.identifier.orcid0000-0002-2384-9141-
dc.identifier.orcid0000-0002-9613-1659-
Appears in Collections:DTCYC - Artículos

Files in This Item:
File Description SizeFormat 
TGRS_2018_2860125.pdf1,78 MBAdobe PDFView/Open


This item is licensed under a Creative Commons License Creative Commons