Learning Priors of Human Motion with Vision Transformers

A clear understanding of where humans move in a scenario, their usual paths and speeds, and where they stop, is very important for different applications, such as mobility studies in urban areas or robot navigation tasks within human-populated environments. We propose in this article, a neural archi...

ver descrição completa

Detalhes bibliográficos
Autores: Falqueto, Placido, Sanfeliu, Alberto, Palopoli, Luigi, Fontanelli, Daniele
Formato: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2024
País:España
Recursos:Consejo Superior de Investigaciones Científicas (CSIC)
Repositorio:DIGITAL.CSIC. Repositorio Institucional del CSIC
OAI Identifier:oai:digital.csic.es:10261/388045
Acesso em linha:http://hdl.handle.net/10261/388045
https://api.elsevier.com/content/abstract/scopus_id/85204095548
Access Level:acceso abierto
Palavra-chave:Human motion prediction
Masked autoencoders
Occupancy priors
Semantic scene understanding
Vision transformers
id ES_905bb5688a9f0187e0efb01aa76b3fcf
oai_identifier_str oai:digital.csic.es:10261/388045
network_acronym_str ES
network_name_str España
repository_id_str
spelling Learning Priors of Human Motion with Vision TransformersFalqueto, PlacidoSanfeliu, AlbertoPalopoli, LuigiFontanelli, DanieleHuman motion predictionMasked autoencodersOccupancy priorsSemantic scene understandingVision transformersA clear understanding of where humans move in a scenario, their usual paths and speeds, and where they stop, is very important for different applications, such as mobility studies in urban areas or robot navigation tasks within human-populated environments. We propose in this article, a neural architecture based on Vision Transformers (ViTs) to provide this information. This solution can arguably capture spatial correlations more effectively than Convolutional Neural Networks (CNNs). In the paper, we describe the methodology and proposed neural architecture and show the experiments' results with a standard dataset. We show that the proposed ViT architecture improves the metrics compared to a method based on a CNN.Peer reviewedInstitute of Electrical and Electronics EngineersSanfeliu, Alberto [0000-0003-3868-9678]Consejo Superior de Investigaciones Científicas [https://ror.org/02gfc7t72]202520252024info:eu-repo/semantics/articlehttp://purl.org/coar/resource_type/c_6501Postprintinfo:eu-repo/semantics/acceptedVersionapplication/pdfhttp://hdl.handle.net/10261/388045https://api.elsevier.com/content/abstract/scopus_id/85204095548reponame:DIGITAL.CSIC. Repositorio Institucional del CSICinstname:Consejo Superior de Investigaciones Científicas (CSIC)Ingléshttps://doi.org/10.1109/COMPSAC61105.2024.00060Síinfo:eu-repo/semantics/openAccessoai:digital.csic.es:10261/3880452026-05-22T06:33:51Z
dc.title.none.fl_str_mv Learning Priors of Human Motion with Vision Transformers
title Learning Priors of Human Motion with Vision Transformers
spellingShingle Learning Priors of Human Motion with Vision Transformers
Falqueto, Placido
Human motion prediction
Masked autoencoders
Occupancy priors
Semantic scene understanding
Vision transformers
title_short Learning Priors of Human Motion with Vision Transformers
title_full Learning Priors of Human Motion with Vision Transformers
title_fullStr Learning Priors of Human Motion with Vision Transformers
title_full_unstemmed Learning Priors of Human Motion with Vision Transformers
title_sort Learning Priors of Human Motion with Vision Transformers
dc.creator.none.fl_str_mv Falqueto, Placido
Sanfeliu, Alberto
Palopoli, Luigi
Fontanelli, Daniele
author Falqueto, Placido
author_facet Falqueto, Placido
Sanfeliu, Alberto
Palopoli, Luigi
Fontanelli, Daniele
author_role author
author2 Sanfeliu, Alberto
Palopoli, Luigi
Fontanelli, Daniele
author2_role author
author
author
dc.contributor.none.fl_str_mv Sanfeliu, Alberto [0000-0003-3868-9678]
Consejo Superior de Investigaciones Científicas [https://ror.org/02gfc7t72]
dc.subject.none.fl_str_mv Human motion prediction
Masked autoencoders
Occupancy priors
Semantic scene understanding
Vision transformers
topic Human motion prediction
Masked autoencoders
Occupancy priors
Semantic scene understanding
Vision transformers
description A clear understanding of where humans move in a scenario, their usual paths and speeds, and where they stop, is very important for different applications, such as mobility studies in urban areas or robot navigation tasks within human-populated environments. We propose in this article, a neural architecture based on Vision Transformers (ViTs) to provide this information. This solution can arguably capture spatial correlations more effectively than Convolutional Neural Networks (CNNs). In the paper, we describe the methodology and proposed neural architecture and show the experiments' results with a standard dataset. We show that the proposed ViT architecture improves the metrics compared to a method based on a CNN.
publishDate 2024
dc.date.none.fl_str_mv 2024
2025
2025
dc.type.none.fl_str_mv info:eu-repo/semantics/article
http://purl.org/coar/resource_type/c_6501
Postprint
info:eu-repo/semantics/acceptedVersion
format article
status_str acceptedVersion
dc.identifier.none.fl_str_mv http://hdl.handle.net/10261/388045
https://api.elsevier.com/content/abstract/scopus_id/85204095548
url http://hdl.handle.net/10261/388045
https://api.elsevier.com/content/abstract/scopus_id/85204095548
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.relation.none.fl_str_mv https://doi.org/10.1109/COMPSAC61105.2024.00060
Sí
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Institute of Electrical and Electronics Engineers
publisher.none.fl_str_mv Institute of Electrical and Electronics Engineers
dc.source.none.fl_str_mv reponame:DIGITAL.CSIC. Repositorio Institucional del CSIC
instname:Consejo Superior de Investigaciones Científicas (CSIC)
instname_str Consejo Superior de Investigaciones Científicas (CSIC)
reponame_str DIGITAL.CSIC. Repositorio Institucional del CSIC
collection DIGITAL.CSIC. Repositorio Institucional del CSIC
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869413287004733440
score 15,812429

Registos relacionados