GeneGAN: Learning Object Transfiguration and Attribute Subspace from Unpaired Data

Object Transfiguration replaces an object in an image with another object from a second image. For example it can perform tasks like "putting exactly those eyeglasses from image A on the nose of the person in image B". Usage of exemplar images allows more precise specification of desired m...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2017-05
Hauptverfasser: Zhou, Shuchang, Xiao, Taihong, Yang, Yi, Feng, Dieqiao, He, Qinyao, He, Weiran
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 Zhou, Shuchang
Xiao, Taihong
Yang, Yi
Feng, Dieqiao
He, Qinyao
He, Weiran
description Object Transfiguration replaces an object in an image with another object from a second image. For example it can perform tasks like "putting exactly those eyeglasses from image A on the nose of the person in image B". Usage of exemplar images allows more precise specification of desired modifications and improves the diversity of conditional image generation. However, previous methods that rely on feature space operations, require paired data and/or appearance models for training or disentangling objects from background. In this work, we propose a model that can learn object transfiguration from two unpaired sets of images: one set containing images that "have" that kind of object, and the other set being the opposite, with the mild constraint that the objects be located approximately at the same place. For example, the training data can be one set of reference face images that have eyeglasses, and another set of images that have not, both of which spatially aligned by face landmarks. Despite the weak 0/1 labels, our model can learn an "eyeglasses" subspace that contain multiple representatives of different types of glasses. Consequently, we can perform fine-grained control of generated images, like swapping the glasses in two images by swapping the projected components in the "eyeglasses" subspace, to create novel images of people wearing eyeglasses. Overall, our deterministic generative model learns disentangled attribute subspaces from weakly labeled data by adversarial training. Experiments on CelebA and Multi-PIE datasets validate the effectiveness of the proposed model on real world data, in generating images with specified eyeglasses, smiling, hair styles, and lighting conditions etc. The code is available online.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2075579788</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2075579788</sourcerecordid><originalsourceid>FETCH-proquest_journals_20755797883</originalsourceid><addsrcrecordid>eNqNyssKgkAUgOEhCJLyHQ60Fmxs0tpJ10UUlK3lqEcZqRmby_vXogdo9S--f8QCniSLKFtyPmGhtX0cx3yVciGSgN2OpOiYXzZwJjRKqg6uVU-1g8Kgsq3svEEntQJUDeTOGVl5R3D3lR2wJmiNfsFDDSgNNbBDhzM2bvFpKfx1yuaHfbE9RYPRb0_Wlb32Rn2p5HEqRLpOsyz57_oAl2Y_RA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2075579788</pqid></control><display><type>article</type><title>GeneGAN: Learning Object Transfiguration and Attribute Subspace from Unpaired Data</title><source>Free E- Journals</source><creator>Zhou, Shuchang ; Xiao, Taihong ; Yang, Yi ; Feng, Dieqiao ; He, Qinyao ; He, Weiran</creator><creatorcontrib>Zhou, Shuchang ; Xiao, Taihong ; Yang, Yi ; Feng, Dieqiao ; He, Qinyao ; He, Weiran</creatorcontrib><description>Object Transfiguration replaces an object in an image with another object from a second image. For example it can perform tasks like "putting exactly those eyeglasses from image A on the nose of the person in image B". Usage of exemplar images allows more precise specification of desired modifications and improves the diversity of conditional image generation. However, previous methods that rely on feature space operations, require paired data and/or appearance models for training or disentangling objects from background. In this work, we propose a model that can learn object transfiguration from two unpaired sets of images: one set containing images that "have" that kind of object, and the other set being the opposite, with the mild constraint that the objects be located approximately at the same place. For example, the training data can be one set of reference face images that have eyeglasses, and another set of images that have not, both of which spatially aligned by face landmarks. Despite the weak 0/1 labels, our model can learn an "eyeglasses" subspace that contain multiple representatives of different types of glasses. Consequently, we can perform fine-grained control of generated images, like swapping the glasses in two images by swapping the projected components in the "eyeglasses" subspace, to create novel images of people wearing eyeglasses. Overall, our deterministic generative model learns disentangled attribute subspaces from weakly labeled data by adversarial training. Experiments on CelebA and Multi-PIE datasets validate the effectiveness of the proposed model on real world data, in generating images with specified eyeglasses, smiling, hair styles, and lighting conditions etc. The code is available online.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Eyeglasses ; Eyewear ; Image processing ; Subspaces ; Training</subject><ispartof>arXiv.org, 2017-05</ispartof><rights>2017. 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>780,784</link.rule.ids></links><search><creatorcontrib>Zhou, Shuchang</creatorcontrib><creatorcontrib>Xiao, Taihong</creatorcontrib><creatorcontrib>Yang, Yi</creatorcontrib><creatorcontrib>Feng, Dieqiao</creatorcontrib><creatorcontrib>He, Qinyao</creatorcontrib><creatorcontrib>He, Weiran</creatorcontrib><title>GeneGAN: Learning Object Transfiguration and Attribute Subspace from Unpaired Data</title><title>arXiv.org</title><description>Object Transfiguration replaces an object in an image with another object from a second image. For example it can perform tasks like "putting exactly those eyeglasses from image A on the nose of the person in image B". Usage of exemplar images allows more precise specification of desired modifications and improves the diversity of conditional image generation. However, previous methods that rely on feature space operations, require paired data and/or appearance models for training or disentangling objects from background. In this work, we propose a model that can learn object transfiguration from two unpaired sets of images: one set containing images that "have" that kind of object, and the other set being the opposite, with the mild constraint that the objects be located approximately at the same place. For example, the training data can be one set of reference face images that have eyeglasses, and another set of images that have not, both of which spatially aligned by face landmarks. Despite the weak 0/1 labels, our model can learn an "eyeglasses" subspace that contain multiple representatives of different types of glasses. Consequently, we can perform fine-grained control of generated images, like swapping the glasses in two images by swapping the projected components in the "eyeglasses" subspace, to create novel images of people wearing eyeglasses. Overall, our deterministic generative model learns disentangled attribute subspaces from weakly labeled data by adversarial training. Experiments on CelebA and Multi-PIE datasets validate the effectiveness of the proposed model on real world data, in generating images with specified eyeglasses, smiling, hair styles, and lighting conditions etc. The code is available online.</description><subject>Eyeglasses</subject><subject>Eyewear</subject><subject>Image processing</subject><subject>Subspaces</subject><subject>Training</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2017</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNqNyssKgkAUgOEhCJLyHQ60Fmxs0tpJ10UUlK3lqEcZqRmby_vXogdo9S--f8QCniSLKFtyPmGhtX0cx3yVciGSgN2OpOiYXzZwJjRKqg6uVU-1g8Kgsq3svEEntQJUDeTOGVl5R3D3lR2wJmiNfsFDDSgNNbBDhzM2bvFpKfx1yuaHfbE9RYPRb0_Wlb32Rn2p5HEqRLpOsyz57_oAl2Y_RA</recordid><startdate>20170514</startdate><enddate>20170514</enddate><creator>Zhou, Shuchang</creator><creator>Xiao, Taihong</creator><creator>Yang, Yi</creator><creator>Feng, Dieqiao</creator><creator>He, Qinyao</creator><creator>He, Weiran</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>20170514</creationdate><title>GeneGAN: Learning Object Transfiguration and Attribute Subspace from Unpaired Data</title><author>Zhou, Shuchang ; Xiao, Taihong ; Yang, Yi ; Feng, Dieqiao ; He, Qinyao ; He, Weiran</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_20755797883</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2017</creationdate><topic>Eyeglasses</topic><topic>Eyewear</topic><topic>Image processing</topic><topic>Subspaces</topic><topic>Training</topic><toplevel>online_resources</toplevel><creatorcontrib>Zhou, Shuchang</creatorcontrib><creatorcontrib>Xiao, Taihong</creatorcontrib><creatorcontrib>Yang, Yi</creatorcontrib><creatorcontrib>Feng, Dieqiao</creatorcontrib><creatorcontrib>He, Qinyao</creatorcontrib><creatorcontrib>He, Weiran</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>Zhou, Shuchang</au><au>Xiao, Taihong</au><au>Yang, Yi</au><au>Feng, Dieqiao</au><au>He, Qinyao</au><au>He, Weiran</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>GeneGAN: Learning Object Transfiguration and Attribute Subspace from Unpaired Data</atitle><jtitle>arXiv.org</jtitle><date>2017-05-14</date><risdate>2017</risdate><eissn>2331-8422</eissn><abstract>Object Transfiguration replaces an object in an image with another object from a second image. For example it can perform tasks like "putting exactly those eyeglasses from image A on the nose of the person in image B". Usage of exemplar images allows more precise specification of desired modifications and improves the diversity of conditional image generation. However, previous methods that rely on feature space operations, require paired data and/or appearance models for training or disentangling objects from background. In this work, we propose a model that can learn object transfiguration from two unpaired sets of images: one set containing images that "have" that kind of object, and the other set being the opposite, with the mild constraint that the objects be located approximately at the same place. For example, the training data can be one set of reference face images that have eyeglasses, and another set of images that have not, both of which spatially aligned by face landmarks. Despite the weak 0/1 labels, our model can learn an "eyeglasses" subspace that contain multiple representatives of different types of glasses. Consequently, we can perform fine-grained control of generated images, like swapping the glasses in two images by swapping the projected components in the "eyeglasses" subspace, to create novel images of people wearing eyeglasses. Overall, our deterministic generative model learns disentangled attribute subspaces from weakly labeled data by adversarial training. Experiments on CelebA and Multi-PIE datasets validate the effectiveness of the proposed model on real world data, in generating images with specified eyeglasses, smiling, hair styles, and lighting conditions etc. The code is available online.</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, 2017-05
issn 2331-8422
language eng
recordid cdi_proquest_journals_2075579788
source Free E- Journals
subjects Eyeglasses
Eyewear
Image processing
Subspaces
Training
title GeneGAN: Learning Object Transfiguration and Attribute Subspace from Unpaired Data
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-28T18%3A14%3A07IST&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=GeneGAN:%20Learning%20Object%20Transfiguration%20and%20Attribute%20Subspace%20from%20Unpaired%20Data&rft.jtitle=arXiv.org&rft.au=Zhou,%20Shuchang&rft.date=2017-05-14&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2075579788%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2075579788&rft_id=info:pmid/&rfr_iscdi=true