Collapse by Conditioning: Training Class-conditional GANs with Limited Data

Class-conditioning offers a direct means to control a Generative Adversarial Network (GAN) based on a discrete input variable. While necessary in many applications, the additional information provided by the class labels could even be expected to benefit the training of the GAN itself. On the contra...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2022-03
Hauptverfasser: Shahbazi, Mohamad, Danelljan, Martin, Danda Pani Paudel, Luc Van Gool
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 Shahbazi, Mohamad
Danelljan, Martin
Danda Pani Paudel
Luc Van Gool
description Class-conditioning offers a direct means to control a Generative Adversarial Network (GAN) based on a discrete input variable. While necessary in many applications, the additional information provided by the class labels could even be expected to benefit the training of the GAN itself. On the contrary, we observe that class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generative ability. Motivated by this observation, we propose a training strategy for class-conditional GANs (cGANs) that effectively prevents the observed mode-collapse by leveraging unconditional learning. Our training strategy starts with an unconditional GAN and gradually injects the class conditioning into the generator and the objective function. The proposed method for training cGANs with limited data results not only in stable training but also in generating high-quality images, thanks to the early-stage exploitation of the shared information across classes. We analyze the observed mode collapse problem in comprehensive experiments on four datasets. Our approach demonstrates outstanding results compared with state-of-the-art methods and established baselines. The code is available at https://github.com/mshahbazi72/transitional-cGAN
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2621120914</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2621120914</sourcerecordid><originalsourceid>FETCH-proquest_journals_26211209143</originalsourceid><addsrcrecordid>eNqNi8kKwjAURYMgWLT_8MB1IXkdHHYSJ1BcdV-ijZoSm9qXIv69Crp3dQ-cc3sswDgW0TRBHLCQqOKcYzbBNI0DtpPOWtWQhuMTpKtL442rTX2ZQ94q8yGQVhFFp59UFjaLA8HD-Cvszc14XcJSeTVi_bOypMPvDtl4vcrlNmpad-80-aJyXfv-U4EZCoF8JpL4v-oFyqc8bw</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2621120914</pqid></control><display><type>article</type><title>Collapse by Conditioning: Training Class-conditional GANs with Limited Data</title><source>Free E- Journals</source><creator>Shahbazi, Mohamad ; Danelljan, Martin ; Danda Pani Paudel ; Luc Van Gool</creator><creatorcontrib>Shahbazi, Mohamad ; Danelljan, Martin ; Danda Pani Paudel ; Luc Van Gool</creatorcontrib><description>Class-conditioning offers a direct means to control a Generative Adversarial Network (GAN) based on a discrete input variable. While necessary in many applications, the additional information provided by the class labels could even be expected to benefit the training of the GAN itself. On the contrary, we observe that class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generative ability. Motivated by this observation, we propose a training strategy for class-conditional GANs (cGANs) that effectively prevents the observed mode-collapse by leveraging unconditional learning. Our training strategy starts with an unconditional GAN and gradually injects the class conditioning into the generator and the objective function. The proposed method for training cGANs with limited data results not only in stable training but also in generating high-quality images, thanks to the early-stage exploitation of the shared information across classes. We analyze the observed mode collapse problem in comprehensive experiments on four datasets. Our approach demonstrates outstanding results compared with state-of-the-art methods and established baselines. The code is available at https://github.com/mshahbazi72/transitional-cGAN</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Generative adversarial networks ; Image quality ; Training</subject><ispartof>arXiv.org, 2022-03</ispartof><rights>2022. This work is published under http://creativecommons.org/licenses/by/4.0/ (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>Shahbazi, Mohamad</creatorcontrib><creatorcontrib>Danelljan, Martin</creatorcontrib><creatorcontrib>Danda Pani Paudel</creatorcontrib><creatorcontrib>Luc Van Gool</creatorcontrib><title>Collapse by Conditioning: Training Class-conditional GANs with Limited Data</title><title>arXiv.org</title><description>Class-conditioning offers a direct means to control a Generative Adversarial Network (GAN) based on a discrete input variable. While necessary in many applications, the additional information provided by the class labels could even be expected to benefit the training of the GAN itself. On the contrary, we observe that class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generative ability. Motivated by this observation, we propose a training strategy for class-conditional GANs (cGANs) that effectively prevents the observed mode-collapse by leveraging unconditional learning. Our training strategy starts with an unconditional GAN and gradually injects the class conditioning into the generator and the objective function. The proposed method for training cGANs with limited data results not only in stable training but also in generating high-quality images, thanks to the early-stage exploitation of the shared information across classes. We analyze the observed mode collapse problem in comprehensive experiments on four datasets. Our approach demonstrates outstanding results compared with state-of-the-art methods and established baselines. The code is available at https://github.com/mshahbazi72/transitional-cGAN</description><subject>Generative adversarial networks</subject><subject>Image quality</subject><subject>Training</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNqNi8kKwjAURYMgWLT_8MB1IXkdHHYSJ1BcdV-ijZoSm9qXIv69Crp3dQ-cc3sswDgW0TRBHLCQqOKcYzbBNI0DtpPOWtWQhuMTpKtL442rTX2ZQ94q8yGQVhFFp59UFjaLA8HD-Cvszc14XcJSeTVi_bOypMPvDtl4vcrlNmpad-80-aJyXfv-U4EZCoF8JpL4v-oFyqc8bw</recordid><startdate>20220316</startdate><enddate>20220316</enddate><creator>Shahbazi, Mohamad</creator><creator>Danelljan, Martin</creator><creator>Danda Pani Paudel</creator><creator>Luc Van Gool</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>20220316</creationdate><title>Collapse by Conditioning: Training Class-conditional GANs with Limited Data</title><author>Shahbazi, Mohamad ; Danelljan, Martin ; Danda Pani Paudel ; Luc Van Gool</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_26211209143</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Generative adversarial networks</topic><topic>Image quality</topic><topic>Training</topic><toplevel>online_resources</toplevel><creatorcontrib>Shahbazi, Mohamad</creatorcontrib><creatorcontrib>Danelljan, Martin</creatorcontrib><creatorcontrib>Danda Pani Paudel</creatorcontrib><creatorcontrib>Luc Van Gool</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>Access via ProQuest (Open Access)</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>Shahbazi, Mohamad</au><au>Danelljan, Martin</au><au>Danda Pani Paudel</au><au>Luc Van Gool</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Collapse by Conditioning: Training Class-conditional GANs with Limited Data</atitle><jtitle>arXiv.org</jtitle><date>2022-03-16</date><risdate>2022</risdate><eissn>2331-8422</eissn><abstract>Class-conditioning offers a direct means to control a Generative Adversarial Network (GAN) based on a discrete input variable. While necessary in many applications, the additional information provided by the class labels could even be expected to benefit the training of the GAN itself. On the contrary, we observe that class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generative ability. Motivated by this observation, we propose a training strategy for class-conditional GANs (cGANs) that effectively prevents the observed mode-collapse by leveraging unconditional learning. Our training strategy starts with an unconditional GAN and gradually injects the class conditioning into the generator and the objective function. The proposed method for training cGANs with limited data results not only in stable training but also in generating high-quality images, thanks to the early-stage exploitation of the shared information across classes. We analyze the observed mode collapse problem in comprehensive experiments on four datasets. Our approach demonstrates outstanding results compared with state-of-the-art methods and established baselines. The code is available at https://github.com/mshahbazi72/transitional-cGAN</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, 2022-03
issn 2331-8422
language eng
recordid cdi_proquest_journals_2621120914
source Free E- Journals
subjects Generative adversarial networks
Image quality
Training
title Collapse by Conditioning: Training Class-conditional GANs with Limited Data
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-23T08%3A17%3A34IST&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=Collapse%20by%20Conditioning:%20Training%20Class-conditional%20GANs%20with%20Limited%20Data&rft.jtitle=arXiv.org&rft.au=Shahbazi,%20Mohamad&rft.date=2022-03-16&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2621120914%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2621120914&rft_id=info:pmid/&rfr_iscdi=true