Learning Controllable Disentangled Representations with Decorrelation Regularization

A crucial problem in learning disentangled image representations is controlling the degree of disentanglement during image editing, while preserving the identity of objects. In this work, we propose a simple yet effective model with the encoder-decoder architecture to address this challenge. To enco...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2019-12
Hauptverfasser: Song, Zengjie, Koyejo, Oluwasanmi, Zhang, Jiangshe
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Song, Zengjie
Koyejo, Oluwasanmi
Zhang, Jiangshe
description A crucial problem in learning disentangled image representations is controlling the degree of disentanglement during image editing, while preserving the identity of objects. In this work, we propose a simple yet effective model with the encoder-decoder architecture to address this challenge. To encourage disentanglement, we devise a distance covariance based decorrelation regularization. Further, for the reconstruction step, our model leverages a soft target representation combined with the latent image code. By exploiting the real-valued space of the soft target representations, we are able to synthesize novel images with the designated properties. We also design a classification based protocol to quantitatively evaluate the disentanglement strength of our model. Experimental results show that the proposed model competently disentangles factors of variation, and is able to manipulate face images to synthesize the desired attributes.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2331355703</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2331355703</sourcerecordid><originalsourceid>FETCH-proquest_journals_23313557033</originalsourceid><addsrcrecordid>eNqNjkEKwjAURIMgWLR3CLguxMRY963iwpV0X2L91pRPUn9SBE9vLR7A1fBm3mJmLJFKbbL9VsoFS0PohBByl0utVcKqMxhy1rW88C6SRzRXBF7aAC4a1yLc-AV6gomj9S7wl40PXkLjiQCnblTaAQ3Z94QrNr8bDJD-csnWx0NVnLKe_HOAEOvOD-TGqf5eU1rnQqn_rA8yKUHZ</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2331355703</pqid></control><display><type>article</type><title>Learning Controllable Disentangled Representations with Decorrelation Regularization</title><source>Free E- Journals</source><creator>Song, Zengjie ; Koyejo, Oluwasanmi ; Zhang, Jiangshe</creator><creatorcontrib>Song, Zengjie ; Koyejo, Oluwasanmi ; Zhang, Jiangshe</creatorcontrib><description>A crucial problem in learning disentangled image representations is controlling the degree of disentanglement during image editing, while preserving the identity of objects. In this work, we propose a simple yet effective model with the encoder-decoder architecture to address this challenge. To encourage disentanglement, we devise a distance covariance based decorrelation regularization. Further, for the reconstruction step, our model leverages a soft target representation combined with the latent image code. By exploiting the real-valued space of the soft target representations, we are able to synthesize novel images with the designated properties. We also design a classification based protocol to quantitatively evaluate the disentanglement strength of our model. Experimental results show that the proposed model competently disentangles factors of variation, and is able to manipulate face images to synthesize the desired attributes.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Coders ; Covariance ; Encoders-Decoders ; Image classification ; Image manipulation ; Image reconstruction ; Learning ; Regularization ; Representations ; Stability ; Synthesis</subject><ispartof>arXiv.org, 2019-12</ispartof><rights>2019. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</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>776,780</link.rule.ids></links><search><creatorcontrib>Song, Zengjie</creatorcontrib><creatorcontrib>Koyejo, Oluwasanmi</creatorcontrib><creatorcontrib>Zhang, Jiangshe</creatorcontrib><title>Learning Controllable Disentangled Representations with Decorrelation Regularization</title><title>arXiv.org</title><description>A crucial problem in learning disentangled image representations is controlling the degree of disentanglement during image editing, while preserving the identity of objects. In this work, we propose a simple yet effective model with the encoder-decoder architecture to address this challenge. To encourage disentanglement, we devise a distance covariance based decorrelation regularization. Further, for the reconstruction step, our model leverages a soft target representation combined with the latent image code. By exploiting the real-valued space of the soft target representations, we are able to synthesize novel images with the designated properties. We also design a classification based protocol to quantitatively evaluate the disentanglement strength of our model. Experimental results show that the proposed model competently disentangles factors of variation, and is able to manipulate face images to synthesize the desired attributes.</description><subject>Coders</subject><subject>Covariance</subject><subject>Encoders-Decoders</subject><subject>Image classification</subject><subject>Image manipulation</subject><subject>Image reconstruction</subject><subject>Learning</subject><subject>Regularization</subject><subject>Representations</subject><subject>Stability</subject><subject>Synthesis</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2019</creationdate><recordtype>article</recordtype><sourceid>BENPR</sourceid><recordid>eNqNjkEKwjAURIMgWLR3CLguxMRY963iwpV0X2L91pRPUn9SBE9vLR7A1fBm3mJmLJFKbbL9VsoFS0PohBByl0utVcKqMxhy1rW88C6SRzRXBF7aAC4a1yLc-AV6gomj9S7wl40PXkLjiQCnblTaAQ3Z94QrNr8bDJD-csnWx0NVnLKe_HOAEOvOD-TGqf5eU1rnQqn_rA8yKUHZ</recordid><startdate>20191225</startdate><enddate>20191225</enddate><creator>Song, Zengjie</creator><creator>Koyejo, Oluwasanmi</creator><creator>Zhang, Jiangshe</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20191225</creationdate><title>Learning Controllable Disentangled Representations with Decorrelation Regularization</title><author>Song, Zengjie ; Koyejo, Oluwasanmi ; Zhang, Jiangshe</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_23313557033</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2019</creationdate><topic>Coders</topic><topic>Covariance</topic><topic>Encoders-Decoders</topic><topic>Image classification</topic><topic>Image manipulation</topic><topic>Image reconstruction</topic><topic>Learning</topic><topic>Regularization</topic><topic>Representations</topic><topic>Stability</topic><topic>Synthesis</topic><toplevel>online_resources</toplevel><creatorcontrib>Song, Zengjie</creatorcontrib><creatorcontrib>Koyejo, Oluwasanmi</creatorcontrib><creatorcontrib>Zhang, Jiangshe</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Song, Zengjie</au><au>Koyejo, Oluwasanmi</au><au>Zhang, Jiangshe</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Learning Controllable Disentangled Representations with Decorrelation Regularization</atitle><jtitle>arXiv.org</jtitle><date>2019-12-25</date><risdate>2019</risdate><eissn>2331-8422</eissn><abstract>A crucial problem in learning disentangled image representations is controlling the degree of disentanglement during image editing, while preserving the identity of objects. In this work, we propose a simple yet effective model with the encoder-decoder architecture to address this challenge. To encourage disentanglement, we devise a distance covariance based decorrelation regularization. Further, for the reconstruction step, our model leverages a soft target representation combined with the latent image code. By exploiting the real-valued space of the soft target representations, we are able to synthesize novel images with the designated properties. We also design a classification based protocol to quantitatively evaluate the disentanglement strength of our model. Experimental results show that the proposed model competently disentangles factors of variation, and is able to manipulate face images to synthesize the desired attributes.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2019-12
issn 2331-8422
language eng
recordid cdi_proquest_journals_2331355703
source Free E- Journals
subjects Coders
Covariance
Encoders-Decoders
Image classification
Image manipulation
Image reconstruction
Learning
Regularization
Representations
Stability
Synthesis
title Learning Controllable Disentangled Representations with Decorrelation Regularization
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-09T01%3A35%3A24IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Learning%20Controllable%20Disentangled%20Representations%20with%20Decorrelation%20Regularization&rft.jtitle=arXiv.org&rft.au=Song,%20Zengjie&rft.date=2019-12-25&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2331355703%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2331355703&rft_id=info:pmid/&rfr_iscdi=true