Loading…
A Network Combining CNN and Transformer for Blind Image Super-Resolution
Blind super-resolution (SR) requires not only estimating blur kernel, but also super-resolving low-resolution image based on estimated blur kernel. Most blind SR methods use convolutional neural networks (CNNs) for kernel estimation, which cannot exploit long-range dependency within image domain, th...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Conference Proceeding |
| Language: | English |
| Subjects: | |
| Online Access: | Request full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| cited_by | |
|---|---|
| cites | |
| container_end_page | 398 |
| container_issue | |
| container_start_page | 394 |
| container_title | |
| container_volume | |
| creator | Zhang, Shuhao Li, Zuoyong Teng, Shenghua Zeng, Kun |
| description | Blind super-resolution (SR) requires not only estimating blur kernel, but also super-resolving low-resolution image based on estimated blur kernel. Most blind SR methods use convolutional neural networks (CNNs) for kernel estimation, which cannot exploit long-range dependency within image domain, thus failing to predict blur kernel accurately. To address this issue, we propose a network combining CNN and transformer named NCCT for kernel estimation. By modeling local and non-local image priors simultaneously, NCCT outperforms other blind SR methods in terms of kernel estimation accuracy. Moreover, we design a network module named RRFDB for constructing lightweight blind SR network, which runs faster and achieves comparative SR performance with fewer parameters compared with other state-of-the-art blind SR methods. |
| doi_str_mv | 10.1109/ITME56794.2022.00091 |
| format | conference_proceeding |
| fullrecord | <record><control><sourceid>ieee_CHZPO</sourceid><recordid>TN_cdi_ieee_primary_10086404</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10086404</ieee_id><sourcerecordid>10086404</sourcerecordid><originalsourceid>FETCH-LOGICAL-i119t-5b9827a6a9c288495020160ab748e0972665d56257accdfffd498f8b1adb1d3a3</originalsourceid><addsrcrecordid>eNotjl1LwzAYRqMgOGb_wS7yB1rffCeXs0xXmBW0Xo-0SUe0HyPtEP-9hXl14OHwcBDaEMgIAfNYVK87IZXhGQVKMwAw5AYlRhnNBDACRLBbtKJc8ZRpqu9RMk1fi0YkBUnlCu23uPTzzxi_cT72dRjCcMJ5WWI7OFxFO0ztGHsf8QL81IVlLXp78vjjcvYxfffT2F3mMA4P6K613eSTf67R5_Ouyvfp4e2lyLeHNBBi5lTURlNlpTUN1ZobAXSJAVsrrj0YRaUUTkgqlG0a17at40a3uibW1cQxy9Zoc_0N3vvjOYbext8jAdCSA2d_YTBMvg</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>A Network Combining CNN and Transformer for Blind Image Super-Resolution</title><source>IEEE Xplore All Conference Series</source><creator>Zhang, Shuhao ; Li, Zuoyong ; Teng, Shenghua ; Zeng, Kun</creator><creatorcontrib>Zhang, Shuhao ; Li, Zuoyong ; Teng, Shenghua ; Zeng, Kun</creatorcontrib><description>Blind super-resolution (SR) requires not only estimating blur kernel, but also super-resolving low-resolution image based on estimated blur kernel. Most blind SR methods use convolutional neural networks (CNNs) for kernel estimation, which cannot exploit long-range dependency within image domain, thus failing to predict blur kernel accurately. To address this issue, we propose a network combining CNN and transformer named NCCT for kernel estimation. By modeling local and non-local image priors simultaneously, NCCT outperforms other blind SR methods in terms of kernel estimation accuracy. Moreover, we design a network module named RRFDB for constructing lightweight blind SR network, which runs faster and achieves comparative SR performance with fewer parameters compared with other state-of-the-art blind SR methods.</description><identifier>EISSN: 2474-3828</identifier><identifier>EISBN: 9798350310153</identifier><identifier>DOI: 10.1109/ITME56794.2022.00091</identifier><identifier>CODEN: IEEPAD</identifier><language>eng</language><publisher>IEEE</publisher><subject>Blind Super-Resolution ; CNN ; Convolutional neural networks ; Education ; Estimation ; Feature extraction ; Kernel ; Kernel Estimation ; Superresolution ; Transformer ; Transformers</subject><ispartof>2022 12th International Conference on Information Technology in Medicine and Education (ITME)v, 2022, p.394-398</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10086404$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,777,781,786,787,27906,54536,54913</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/10086404$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Zhang, Shuhao</creatorcontrib><creatorcontrib>Li, Zuoyong</creatorcontrib><creatorcontrib>Teng, Shenghua</creatorcontrib><creatorcontrib>Zeng, Kun</creatorcontrib><title>A Network Combining CNN and Transformer for Blind Image Super-Resolution</title><title>2022 12th International Conference on Information Technology in Medicine and Education (ITME)v</title><addtitle>ITME</addtitle><description>Blind super-resolution (SR) requires not only estimating blur kernel, but also super-resolving low-resolution image based on estimated blur kernel. Most blind SR methods use convolutional neural networks (CNNs) for kernel estimation, which cannot exploit long-range dependency within image domain, thus failing to predict blur kernel accurately. To address this issue, we propose a network combining CNN and transformer named NCCT for kernel estimation. By modeling local and non-local image priors simultaneously, NCCT outperforms other blind SR methods in terms of kernel estimation accuracy. Moreover, we design a network module named RRFDB for constructing lightweight blind SR network, which runs faster and achieves comparative SR performance with fewer parameters compared with other state-of-the-art blind SR methods.</description><subject>Blind Super-Resolution</subject><subject>CNN</subject><subject>Convolutional neural networks</subject><subject>Education</subject><subject>Estimation</subject><subject>Feature extraction</subject><subject>Kernel</subject><subject>Kernel Estimation</subject><subject>Superresolution</subject><subject>Transformer</subject><subject>Transformers</subject><issn>2474-3828</issn><isbn>9798350310153</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2022</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNotjl1LwzAYRqMgOGb_wS7yB1rffCeXs0xXmBW0Xo-0SUe0HyPtEP-9hXl14OHwcBDaEMgIAfNYVK87IZXhGQVKMwAw5AYlRhnNBDACRLBbtKJc8ZRpqu9RMk1fi0YkBUnlCu23uPTzzxi_cT72dRjCcMJ5WWI7OFxFO0ztGHsf8QL81IVlLXp78vjjcvYxfffT2F3mMA4P6K613eSTf67R5_Ouyvfp4e2lyLeHNBBi5lTURlNlpTUN1ZobAXSJAVsrrj0YRaUUTkgqlG0a17at40a3uibW1cQxy9Zoc_0N3vvjOYbext8jAdCSA2d_YTBMvg</recordid><startdate>202211</startdate><enddate>202211</enddate><creator>Zhang, Shuhao</creator><creator>Li, Zuoyong</creator><creator>Teng, Shenghua</creator><creator>Zeng, Kun</creator><general>IEEE</general><scope>6IE</scope><scope>6IL</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIL</scope></search><sort><creationdate>202211</creationdate><title>A Network Combining CNN and Transformer for Blind Image Super-Resolution</title><author>Zhang, Shuhao ; Li, Zuoyong ; Teng, Shenghua ; Zeng, Kun</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i119t-5b9827a6a9c288495020160ab748e0972665d56257accdfffd498f8b1adb1d3a3</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Blind Super-Resolution</topic><topic>CNN</topic><topic>Convolutional neural networks</topic><topic>Education</topic><topic>Estimation</topic><topic>Feature extraction</topic><topic>Kernel</topic><topic>Kernel Estimation</topic><topic>Superresolution</topic><topic>Transformer</topic><topic>Transformers</topic><toplevel>online_resources</toplevel><creatorcontrib>Zhang, Shuhao</creatorcontrib><creatorcontrib>Li, Zuoyong</creatorcontrib><creatorcontrib>Teng, Shenghua</creatorcontrib><creatorcontrib>Zeng, Kun</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Electronic Library (IEL)</collection><collection>IEEE Proceedings Order Plans (POP All) 1998-Present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Zhang, Shuhao</au><au>Li, Zuoyong</au><au>Teng, Shenghua</au><au>Zeng, Kun</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>A Network Combining CNN and Transformer for Blind Image Super-Resolution</atitle><btitle>2022 12th International Conference on Information Technology in Medicine and Education (ITME)v</btitle><stitle>ITME</stitle><date>2022-11</date><risdate>2022</risdate><spage>394</spage><epage>398</epage><pages>394-398</pages><eissn>2474-3828</eissn><eisbn>9798350310153</eisbn><coden>IEEPAD</coden><abstract>Blind super-resolution (SR) requires not only estimating blur kernel, but also super-resolving low-resolution image based on estimated blur kernel. Most blind SR methods use convolutional neural networks (CNNs) for kernel estimation, which cannot exploit long-range dependency within image domain, thus failing to predict blur kernel accurately. To address this issue, we propose a network combining CNN and transformer named NCCT for kernel estimation. By modeling local and non-local image priors simultaneously, NCCT outperforms other blind SR methods in terms of kernel estimation accuracy. Moreover, we design a network module named RRFDB for constructing lightweight blind SR network, which runs faster and achieves comparative SR performance with fewer parameters compared with other state-of-the-art blind SR methods.</abstract><pub>IEEE</pub><doi>10.1109/ITME56794.2022.00091</doi><tpages>5</tpages></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | EISSN: 2474-3828 |
| ispartof | 2022 12th International Conference on Information Technology in Medicine and Education (ITME)v, 2022, p.394-398 |
| issn | 2474-3828 |
| language | eng |
| recordid | cdi_ieee_primary_10086404 |
| source | IEEE Xplore All Conference Series |
| subjects | Blind Super-Resolution CNN Convolutional neural networks Education Estimation Feature extraction Kernel Kernel Estimation Superresolution Transformer Transformers |
| title | A Network Combining CNN and Transformer for Blind Image Super-Resolution |
| url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-19T21%3A57%3A44IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=A%20Network%20Combining%20CNN%20and%20Transformer%20for%20Blind%20Image%20Super-Resolution&rft.btitle=2022%2012th%20International%20Conference%20on%20Information%20Technology%20in%20Medicine%20and%20Education%20(ITME)v&rft.au=Zhang,%20Shuhao&rft.date=2022-11&rft.spage=394&rft.epage=398&rft.pages=394-398&rft.eissn=2474-3828&rft.coden=IEEPAD&rft_id=info:doi/10.1109/ITME56794.2022.00091&rft.eisbn=9798350310153&rft_dat=%3Cieee_CHZPO%3E10086404%3C/ieee_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i119t-5b9827a6a9c288495020160ab748e0972665d56257accdfffd498f8b1adb1d3a3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=10086404&rfr_iscdi=true |