Repository logo
 
Publication

Human Pose Estimation by a Series of Residual Auto-Encoders

2017Conference object Restricted access
Simple item page
dc.contributor.authorFarrajota, Miguel
dc.contributor.authorRodrigues, João
dc.contributor.authordu Buf, Hans
dc.contributor.editorAlexandre, L. A.
dc.contributor.editorSanchez, J. S.
dc.contributor.editorRodrigues, J. M. F.
dc.date.accessioned2019-11-20T15:07:54Z
dc.date.available2019-11-20T15:07:54Z
dc.date.issued2017
dc.description.abstractPose estimation is the task of predicting the pose of an object in an image or in a sequence of images. Here, we focus on articulated human pose estimation in scenes with a single person. We employ a series of residual auto-encoders to produce multiple predictions which are then combined to provide a heatmap prediction of body joints. In this network topology, features are processed across all scales which captures the various spatial relationships associated with the body. Repeated bottom-up and top-down processing with intermediate supervision for each auto-encoder network is applied. We propose some improvements to this type of regression-based networks to further increase performance, namely: (a) increase the number of parameters of the auto-encoder networks in the pipeline, (b) use stronger regularization along with heavy data augmentation, (c) use sub-pixel precision for more precise joint localization, and (d) combine all auto-encoders output heatmaps into a single prediction, which further increases body joint prediction accuracy. We demonstrate state-of-the-art results on the popular FLIC and LSP datasets.
dc.description.sponsorshipFCT project LARSyS [UID/EEA/50009/2013]
dc.description.sponsorshipFCT PhD grant [SFRH/BD/79812/2011]
dc.identifier.doi10.1007/978-3-319-58838-4_15
dc.identifier.isbn978-3-319-58838-4
dc.identifier.isbn978-3-319-58837-7
dc.identifier.issn0302-9743
dc.identifier.issn1611-3349
dc.identifier.urihttp://hdl.handle.net/10400.1/13271
dc.language.isoeng
dc.peerreviewedyes
dc.publisherSpringer International Publishing Ag
dc.relationNEURAL CORRELATES OF MOTION AND STEREO VISION IN HUMAN POSE AND GAIT DETECTION
dc.relation.ispartofseriesLecture Notes in Computer Science
dc.titleHuman Pose Estimation by a Series of Residual Auto-Encoders
dc.typeconference object
dspace.entity.typePublication
oaire.awardTitleNEURAL CORRELATES OF MOTION AND STEREO VISION IN HUMAN POSE AND GAIT DETECTION
oaire.awardURIinfo:eu-repo/grantAgreement/FCT/5876/UID%2FEEA%2F50009%2F2013/PT
oaire.awardURIinfo:eu-repo/grantAgreement/FCT//SFRH%2FBD%2F79812%2F2011/PT
oaire.citation.conferencePlaceUniv Algarve, Faro, PORTUGAL
oaire.citation.endPage139
oaire.citation.startPage131
oaire.citation.titlePattern Recognition and Image Analysis (Ibpria 2017)
oaire.citation.title8th Iberian Conference on Pattern Recognition and Image Analysis (Ibpria)
oaire.citation.volume10255
oaire.fundingStream5876
person.familyNameFarrajota
person.familyNameRodrigues
person.familyNamedu Buf
person.givenNameMiguel
person.givenNameJoao
person.givenNameHans
person.identifier.ciencia-id8A19-98F7-9914
person.identifier.orcid0000-0001-7970-4649
person.identifier.orcid0000-0002-3562-6025
person.identifier.orcid0000-0002-4345-1237
person.identifier.ridM-5125-2013
person.identifier.scopus-author-id55807461600
person.identifier.scopus-author-id6604075916
project.funder.identifierhttp://doi.org/10.13039/501100001871
project.funder.identifierhttp://doi.org/10.13039/501100001871
project.funder.nameFundação para a Ciência e a Tecnologia
project.funder.nameFundação para a Ciência e a Tecnologia
rcaap.rightsrestrictedAccess
rcaap.typeconferenceObject
relation.isAuthorOfPublicationca40db25-2109-4390-ad94-c5f28803c7e8
relation.isAuthorOfPublication683ba85b-459c-4789-a4ff-a4e2a904b295
relation.isAuthorOfPublicationcfad5636-2c77-4db0-a3a0-d7eb97ce6bee
relation.isAuthorOfPublication.latestForDiscovery683ba85b-459c-4789-a4ff-a4e2a904b295
relation.isProjectOfPublication19a727e2-d775-407f-ab10-7d5f19577e08
relation.isProjectOfPublicationd150d21b-0a3e-4403-92cb-65f221ecb683
relation.isProjectOfPublication.latestForDiscovery19a727e2-d775-407f-ab10-7d5f19577e08

Files

Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
13271.pdf
Size:
952.55 KB
Format:
Adobe Portable Document Format
Download

Collections