Learning Pose Specific Representations by Predicting Different Views

The labeled data required to learn pose estimation for articulated objects is difficult to provide in the desired quantity, realism, density, and accuracy. To address this issue, we develop a method to learn representations, which are very specific for articulated poses, without the need for labeled...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Poier, Georg, Schinagl, David, Bischof, Horst
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Poier, Georg
Schinagl, David
Bischof, Horst
description The labeled data required to learn pose estimation for articulated objects is difficult to provide in the desired quantity, realism, density, and accuracy. To address this issue, we develop a method to learn representations, which are very specific for articulated poses, without the need for labeled training data. We exploit the observation that the object pose of a known object is predictive for the appearance in any known view. That is, given only the pose and shape parameters of a hand, the hand's appearance from any viewpoint can be approximated. To exploit this observation, we train a model that -- given input from one view -- estimates a latent representation, which is trained to be predictive for the appearance of the object when captured from another viewpoint. Thus, the only necessary supervision is the second view. The training process of this model reveals an implicit pose representation in the latent space. Importantly, at test time the pose representation can be inferred using only a single view. In qualitative and quantitative experiments we show that the learned representations capture detailed pose information. Moreover, when training the proposed method jointly with labeled and unlabeled data, it consistently surpasses the performance of its fully supervised counterpart, while reducing the amount of needed labeled samples by at least one order of magnitude.
doi_str_mv 10.48550/arxiv.1804.03390
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_1804_03390</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>1804_03390</sourcerecordid><originalsourceid>FETCH-LOGICAL-a670-e21f3dfd83c7317fc4281bec3eea0acb49faf566556c32a6cd6bf415337590123</originalsourceid><addsrcrecordid>eNotz8FOwzAQBFBfekBtP4AT_oEEO2s7yRG1FJAiUUHVa7RxditLkEZ2BPTvoaWnOcxopCfErVa5qaxV9xh_wleuK2VyBVCrG7FuCOMQhoPcHhPJ95F84ODlG42REg0TTuE4JNmd5DZSH_x03q4DM8W_Vu4DfaeFmDF-JFpecy52m8fd6jlrXp9eVg9Nhq5UGRWaoee-Al-CLtmbotIdeSBChb4zNSNb56x1Hgp0vncdG20BSlsrXcBc3P3fXhjtGMMnxlN75rQXDvwCotZFtQ</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Learning Pose Specific Representations by Predicting Different Views</title><source>arXiv.org</source><creator>Poier, Georg ; Schinagl, David ; Bischof, Horst</creator><creatorcontrib>Poier, Georg ; Schinagl, David ; Bischof, Horst</creatorcontrib><description>The labeled data required to learn pose estimation for articulated objects is difficult to provide in the desired quantity, realism, density, and accuracy. To address this issue, we develop a method to learn representations, which are very specific for articulated poses, without the need for labeled training data. We exploit the observation that the object pose of a known object is predictive for the appearance in any known view. That is, given only the pose and shape parameters of a hand, the hand's appearance from any viewpoint can be approximated. To exploit this observation, we train a model that -- given input from one view -- estimates a latent representation, which is trained to be predictive for the appearance of the object when captured from another viewpoint. Thus, the only necessary supervision is the second view. The training process of this model reveals an implicit pose representation in the latent space. Importantly, at test time the pose representation can be inferred using only a single view. In qualitative and quantitative experiments we show that the learned representations capture detailed pose information. Moreover, when training the proposed method jointly with labeled and unlabeled data, it consistently surpasses the performance of its fully supervised counterpart, while reducing the amount of needed labeled samples by at least one order of magnitude.</description><identifier>DOI: 10.48550/arxiv.1804.03390</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition ; Computer Science - Learning</subject><creationdate>2018-04</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/1804.03390$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.1804.03390$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Poier, Georg</creatorcontrib><creatorcontrib>Schinagl, David</creatorcontrib><creatorcontrib>Bischof, Horst</creatorcontrib><title>Learning Pose Specific Representations by Predicting Different Views</title><description>The labeled data required to learn pose estimation for articulated objects is difficult to provide in the desired quantity, realism, density, and accuracy. To address this issue, we develop a method to learn representations, which are very specific for articulated poses, without the need for labeled training data. We exploit the observation that the object pose of a known object is predictive for the appearance in any known view. That is, given only the pose and shape parameters of a hand, the hand's appearance from any viewpoint can be approximated. To exploit this observation, we train a model that -- given input from one view -- estimates a latent representation, which is trained to be predictive for the appearance of the object when captured from another viewpoint. Thus, the only necessary supervision is the second view. The training process of this model reveals an implicit pose representation in the latent space. Importantly, at test time the pose representation can be inferred using only a single view. In qualitative and quantitative experiments we show that the learned representations capture detailed pose information. Moreover, when training the proposed method jointly with labeled and unlabeled data, it consistently surpasses the performance of its fully supervised counterpart, while reducing the amount of needed labeled samples by at least one order of magnitude.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2018</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz8FOwzAQBFBfekBtP4AT_oEEO2s7yRG1FJAiUUHVa7RxditLkEZ2BPTvoaWnOcxopCfErVa5qaxV9xh_wleuK2VyBVCrG7FuCOMQhoPcHhPJ95F84ODlG42REg0TTuE4JNmd5DZSH_x03q4DM8W_Vu4DfaeFmDF-JFpecy52m8fd6jlrXp9eVg9Nhq5UGRWaoee-Al-CLtmbotIdeSBChb4zNSNb56x1Hgp0vncdG20BSlsrXcBc3P3fXhjtGMMnxlN75rQXDvwCotZFtQ</recordid><startdate>20180410</startdate><enddate>20180410</enddate><creator>Poier, Georg</creator><creator>Schinagl, David</creator><creator>Bischof, Horst</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20180410</creationdate><title>Learning Pose Specific Representations by Predicting Different Views</title><author>Poier, Georg ; Schinagl, David ; Bischof, Horst</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a670-e21f3dfd83c7317fc4281bec3eea0acb49faf566556c32a6cd6bf415337590123</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2018</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Poier, Georg</creatorcontrib><creatorcontrib>Schinagl, David</creatorcontrib><creatorcontrib>Bischof, Horst</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Poier, Georg</au><au>Schinagl, David</au><au>Bischof, Horst</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Learning Pose Specific Representations by Predicting Different Views</atitle><date>2018-04-10</date><risdate>2018</risdate><abstract>The labeled data required to learn pose estimation for articulated objects is difficult to provide in the desired quantity, realism, density, and accuracy. To address this issue, we develop a method to learn representations, which are very specific for articulated poses, without the need for labeled training data. We exploit the observation that the object pose of a known object is predictive for the appearance in any known view. That is, given only the pose and shape parameters of a hand, the hand's appearance from any viewpoint can be approximated. To exploit this observation, we train a model that -- given input from one view -- estimates a latent representation, which is trained to be predictive for the appearance of the object when captured from another viewpoint. Thus, the only necessary supervision is the second view. The training process of this model reveals an implicit pose representation in the latent space. Importantly, at test time the pose representation can be inferred using only a single view. In qualitative and quantitative experiments we show that the learned representations capture detailed pose information. Moreover, when training the proposed method jointly with labeled and unlabeled data, it consistently surpasses the performance of its fully supervised counterpart, while reducing the amount of needed labeled samples by at least one order of magnitude.</abstract><doi>10.48550/arxiv.1804.03390</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.1804.03390
ispartof
issn
language eng
recordid cdi_arxiv_primary_1804_03390
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
Computer Science - Learning
title Learning Pose Specific Representations by Predicting Different Views
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-22T14%3A44%3A25IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Learning%20Pose%20Specific%20Representations%20by%20Predicting%20Different%20Views&rft.au=Poier,%20Georg&rft.date=2018-04-10&rft_id=info:doi/10.48550/arxiv.1804.03390&rft_dat=%3Carxiv_GOX%3E1804_03390%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true