Improved SinGAN's Performance by Changing the Activation Function
Generative adversarial nets (GANs) perform well on a variety of tasks, but rely on large datasets and expensive computer‐based learning. SinGAN was proposed as a GANs that overcomes this problem, but its super‐resolution performance for large‐scale natural images was not very good. In this study, we...
Gespeichert in:
| Veröffentlicht in: | IEEJ transactions on electrical and electronic engineering 2022-02, Vol.17 (2), p.308-310 |
|---|---|
| Hauptverfasser: | , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 310 |
|---|---|
| container_issue | 2 |
| container_start_page | 308 |
| container_title | IEEJ transactions on electrical and electronic engineering |
| container_volume | 17 |
| creator | Segawa, Ryo Hayashi, Hitoshi |
| description | Generative adversarial nets (GANs) perform well on a variety of tasks, but rely on large datasets and expensive computer‐based learning. SinGAN was proposed as a GANs that overcomes this problem, but its super‐resolution performance for large‐scale natural images was not very good. In this study, we aimed to improve the performance of SinGAN by changing the activation function. As a result of the verification, it was found that applying RSwish to the generator and Swish to the discriminator can generate an image with less deterioration as a whole than the default. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC. |
| doi_str_mv | 10.1002/tee.23514 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_2617200663</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2617200663</sourcerecordid><originalsourceid>FETCH-LOGICAL-c2974-705fe6d97ce644280f61b5a5b86c9f94badcb153c6f8790c9912c0ac8e986ad33</originalsourceid><addsrcrecordid>eNp1kE9PwjAYhxujiYge_AZNPBgPg_5b1x4XAkhC1EQ8N13Xwgh02A7Mvr3DGW-e3t_h-b3vmweAe4xGGCEybqwdEZpidgEGWFKcMCnw5V_O6DW4iXGLEONUiAHIF_tDqE-2hO-Vn-cvjxG-2eDqsNfeWFi0cLLRfl35NWw2FuamqU66qWoPZ0dvzuEWXDm9i_budw7Bx2y6mjwny9f5YpIvE0NkxpIMpc7yUmbGcsaIQI7jItVpIbiRTrJCl6bAKTXciUwiIyUmBmkjrBRcl5QOwUO_t_v382hjo7b1MfjupCIcZwQhzs_UU0-ZUMcYrFOHUO11aBVG6mxIdYbUj6GOHffsV7Wz7f-gWk2nfeMbVfVmUA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2617200663</pqid></control><display><type>article</type><title>Improved SinGAN's Performance by Changing the Activation Function</title><source>Wiley Online Library Journals Frontfile Complete</source><creator>Segawa, Ryo ; Hayashi, Hitoshi</creator><creatorcontrib>Segawa, Ryo ; Hayashi, Hitoshi</creatorcontrib><description>Generative adversarial nets (GANs) perform well on a variety of tasks, but rely on large datasets and expensive computer‐based learning. SinGAN was proposed as a GANs that overcomes this problem, but its super‐resolution performance for large‐scale natural images was not very good. In this study, we aimed to improve the performance of SinGAN by changing the activation function. As a result of the verification, it was found that applying RSwish to the generator and Swish to the discriminator can generate an image with less deterioration as a whole than the default. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.</description><identifier>ISSN: 1931-4973</identifier><identifier>EISSN: 1931-4981</identifier><identifier>DOI: 10.1002/tee.23514</identifier><language>eng</language><publisher>Hoboken, USA: John Wiley & Sons, Inc</publisher><subject>activation function ; deep learning ; GANs ; Performance enhancement ; super‐resolution ; unsupervised learning</subject><ispartof>IEEJ transactions on electrical and electronic engineering, 2022-02, Vol.17 (2), p.308-310</ispartof><rights>2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.</rights><rights>2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c2974-705fe6d97ce644280f61b5a5b86c9f94badcb153c6f8790c9912c0ac8e986ad33</citedby><cites>FETCH-LOGICAL-c2974-705fe6d97ce644280f61b5a5b86c9f94badcb153c6f8790c9912c0ac8e986ad33</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://onlinelibrary.wiley.com/doi/pdf/10.1002%2Ftee.23514$$EPDF$$P50$$Gwiley$$H</linktopdf><linktohtml>$$Uhttps://onlinelibrary.wiley.com/doi/full/10.1002%2Ftee.23514$$EHTML$$P50$$Gwiley$$H</linktohtml><link.rule.ids>314,776,780,1411,27903,27904,45553,45554</link.rule.ids></links><search><creatorcontrib>Segawa, Ryo</creatorcontrib><creatorcontrib>Hayashi, Hitoshi</creatorcontrib><title>Improved SinGAN's Performance by Changing the Activation Function</title><title>IEEJ transactions on electrical and electronic engineering</title><description>Generative adversarial nets (GANs) perform well on a variety of tasks, but rely on large datasets and expensive computer‐based learning. SinGAN was proposed as a GANs that overcomes this problem, but its super‐resolution performance for large‐scale natural images was not very good. In this study, we aimed to improve the performance of SinGAN by changing the activation function. As a result of the verification, it was found that applying RSwish to the generator and Swish to the discriminator can generate an image with less deterioration as a whole than the default. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.</description><subject>activation function</subject><subject>deep learning</subject><subject>GANs</subject><subject>Performance enhancement</subject><subject>super‐resolution</subject><subject>unsupervised learning</subject><issn>1931-4973</issn><issn>1931-4981</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><recordid>eNp1kE9PwjAYhxujiYge_AZNPBgPg_5b1x4XAkhC1EQ8N13Xwgh02A7Mvr3DGW-e3t_h-b3vmweAe4xGGCEybqwdEZpidgEGWFKcMCnw5V_O6DW4iXGLEONUiAHIF_tDqE-2hO-Vn-cvjxG-2eDqsNfeWFi0cLLRfl35NWw2FuamqU66qWoPZ0dvzuEWXDm9i_budw7Bx2y6mjwny9f5YpIvE0NkxpIMpc7yUmbGcsaIQI7jItVpIbiRTrJCl6bAKTXciUwiIyUmBmkjrBRcl5QOwUO_t_v382hjo7b1MfjupCIcZwQhzs_UU0-ZUMcYrFOHUO11aBVG6mxIdYbUj6GOHffsV7Wz7f-gWk2nfeMbVfVmUA</recordid><startdate>202202</startdate><enddate>202202</enddate><creator>Segawa, Ryo</creator><creator>Hayashi, Hitoshi</creator><general>John Wiley & Sons, Inc</general><general>Wiley Subscription Services, Inc</general><scope>AAYXX</scope><scope>CITATION</scope><scope>7SP</scope><scope>8FD</scope><scope>L7M</scope></search><sort><creationdate>202202</creationdate><title>Improved SinGAN's Performance by Changing the Activation Function</title><author>Segawa, Ryo ; Hayashi, Hitoshi</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c2974-705fe6d97ce644280f61b5a5b86c9f94badcb153c6f8790c9912c0ac8e986ad33</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>activation function</topic><topic>deep learning</topic><topic>GANs</topic><topic>Performance enhancement</topic><topic>super‐resolution</topic><topic>unsupervised learning</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Segawa, Ryo</creatorcontrib><creatorcontrib>Hayashi, Hitoshi</creatorcontrib><collection>CrossRef</collection><collection>Electronics & Communications Abstracts</collection><collection>Technology Research Database</collection><collection>Advanced Technologies Database with Aerospace</collection><jtitle>IEEJ transactions on electrical and electronic engineering</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Segawa, Ryo</au><au>Hayashi, Hitoshi</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Improved SinGAN's Performance by Changing the Activation Function</atitle><jtitle>IEEJ transactions on electrical and electronic engineering</jtitle><date>2022-02</date><risdate>2022</risdate><volume>17</volume><issue>2</issue><spage>308</spage><epage>310</epage><pages>308-310</pages><issn>1931-4973</issn><eissn>1931-4981</eissn><abstract>Generative adversarial nets (GANs) perform well on a variety of tasks, but rely on large datasets and expensive computer‐based learning. SinGAN was proposed as a GANs that overcomes this problem, but its super‐resolution performance for large‐scale natural images was not very good. In this study, we aimed to improve the performance of SinGAN by changing the activation function. As a result of the verification, it was found that applying RSwish to the generator and Swish to the discriminator can generate an image with less deterioration as a whole than the default. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.</abstract><cop>Hoboken, USA</cop><pub>John Wiley & Sons, Inc</pub><doi>10.1002/tee.23514</doi><tpages>3</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1931-4973 |
| ispartof | IEEJ transactions on electrical and electronic engineering, 2022-02, Vol.17 (2), p.308-310 |
| issn | 1931-4973 1931-4981 |
| language | eng |
| recordid | cdi_proquest_journals_2617200663 |
| source | Wiley Online Library Journals Frontfile Complete |
| subjects | activation function deep learning GANs Performance enhancement super‐resolution unsupervised learning |
| title | Improved SinGAN's Performance by Changing the Activation Function |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-21T13%3A57%3A02IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Improved%20SinGAN's%20Performance%20by%20Changing%20the%20Activation%20Function&rft.jtitle=IEEJ%20transactions%20on%20electrical%20and%20electronic%20engineering&rft.au=Segawa,%20Ryo&rft.date=2022-02&rft.volume=17&rft.issue=2&rft.spage=308&rft.epage=310&rft.pages=308-310&rft.issn=1931-4973&rft.eissn=1931-4981&rft_id=info:doi/10.1002/tee.23514&rft_dat=%3Cproquest_cross%3E2617200663%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2617200663&rft_id=info:pmid/&rfr_iscdi=true |