A restricted boltzmann machine-based approach for robust dimensionality reduction

Página do item simplificado
dc.contributor.authorBotelho De Souza, Gustavo
dc.contributor.authorDa Silva Santos, Daniel Felipe [UNESP]
dc.contributor.authorGonalves Pires, Rafael
dc.contributor.authorMarana, Aparecido Nilceu [UNESP]
dc.contributor.authorPapa, Joo Paulo [UNESP]
dc.contributor.institutionUniversidade Federal de São Carlos (UFSCar)
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)
dc.date.accessioned2018-12-11T17:37:52Z
dc.date.available2018-12-11T17:37:52Z
dc.date.issued2018-01-31
dc.description.abstractData dimensionality is an important issue to be adressed by pattern recognition systems. Despite of storage and processing, working with high-dimensional feature vectors also requires complex optimization methods. A proper selection of the most important features is essential and dimensionality reduction techniques can also be applied in order to avoid dealing with more information than needed. One of the most important analytical techniques for such task is Principal Component Analysis (PCA). In this work we propose a novel and more robust dimensionality reduction approach based on the Restricted Boltzmann Machines (RBMs), neural networks able to learn the probability distribution of the set of training samples, identifying the best features to discriminate them, for face spoofing detection. Results of the proposed approach show that the features learned and extracted by RBMs are more robust than the ones analytically obtained by PCA for differentiating between real and fake facial images.en
dc.description.affiliationUFSCar Federal University of Sao Carlos
dc.description.affiliationUNESP Sao Paulo State University
dc.description.affiliationUnespUNESP Sao Paulo State University
dc.format.extent138-143
dc.identifierhttp://dx.doi.org/10.1109/WVC.2017.00031
dc.identifier.citationProceedings - 13th Workshop of Computer Vision, WVC 2017, v. 2018-January, p. 138-143.
dc.identifier.doi10.1109/WVC.2017.00031
dc.identifier.scopus2-s2.0-85050780802
dc.identifier.urihttp://hdl.handle.net/11449/180066
dc.language.isoeng
dc.relation.ispartofProceedings - 13th Workshop of Computer Vision, WVC 2017
dc.rights.accessRightsAcesso aberto
dc.sourceScopus
dc.subjectDimensionality reduction
dc.subjectFace spoofing detection
dc.subjectProbabilistic neural networks
dc.subjectRestricted Boltzmann Machines
dc.titleA restricted boltzmann machine-based approach for robust dimensionality reductionen
dc.typeTrabalho apresentado em evento
unesp.author.lattes6027713750942689[4]
unesp.author.orcid0000-0003-4861-7061[4]
unesp.campusUniversidade Estadual Paulista (Unesp), Faculdade de Ciências, Baurupt
unesp.departmentComputação - FCpt

Arquivos

Coleções

Bauru - FC - Faculdade de Ciências