Does Pooling Really Matter? An Evaluation on Gait Recognition

dc.contributor.authordos Santos, Claudio Filipi Goncalves
dc.contributor.authorMoreira, Thierry Pinheiro [UNESP]
dc.contributor.authorColombo, Danilo
dc.contributor.authorPapa, João Paulo [UNESP]
dc.contributor.institutionUniversidade Federal de São Carlos (UFSCar)
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)
dc.contributor.institutionPetróleo Brasileiro S.A. – Petrobras
dc.date.accessioned2020-12-12T02:30:27Z
dc.date.available2020-12-12T02:30:27Z
dc.date.issued2019-01-01
dc.description.abstractMost Convolutional Neural Networks make use of subsampling layers to reduce dimensionality and keep only the most essential information, besides turning the model more robust to rotation and translation variations. One of the most common sampling methods is the one who keeps only the maximum value in a given region, known as max-pooling. In this study, we provide pieces of evidence that, by removing this subsampling layer and changing the stride of the convolution layer, one can obtain comparable results but much faster. Results on the gait recognition task show the robustness of the proposed approach, as well as its statistical similarity to other pooling methods.en
dc.description.affiliationFederal University of São Carlos - UFSCar
dc.description.affiliationState University of Sao Paulo - UNESP
dc.description.affiliationCenpes Petróleo Brasileiro S.A. – Petrobras
dc.description.affiliationUnespState University of Sao Paulo - UNESP
dc.description.sponsorshipFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.description.sponsorshipConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.description.sponsorshipIdFAPESP: 2013/07375-0
dc.description.sponsorshipIdFAPESP: 2014/12236-1
dc.description.sponsorshipIdFAPESP: 2016/06441-7
dc.description.sponsorshipIdFAPESP: 2017/25908-6
dc.description.sponsorshipIdCNPq: 307066/2017-7
dc.description.sponsorshipIdCNPq: 427968/2018-6
dc.description.sponsorshipIdCNPq: 429003/2018-8
dc.format.extent751-760
dc.identifierhttp://dx.doi.org/10.1007/978-3-030-33904-3_71
dc.identifier.citationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 11896 LNCS, p. 751-760.
dc.identifier.doi10.1007/978-3-030-33904-3_71
dc.identifier.issn1611-3349
dc.identifier.issn0302-9743
dc.identifier.scopus2-s2.0-85075696640
dc.identifier.urihttp://hdl.handle.net/11449/201356
dc.language.isoeng
dc.relation.ispartofLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
dc.sourceScopus
dc.subjectConvolutional Neural Networks
dc.subjectDeep learning
dc.subjectGait recognition
dc.titleDoes Pooling Really Matter? An Evaluation on Gait Recognitionen
dc.typeTrabalho apresentado em evento
unesp.campusUniversidade Estadual Paulista (Unesp), Faculdade de Ciências, Baurupt
unesp.departmentComputação - FCpt

Files