Fechar

1. Identificação
Tipo de ReferênciaArtigo em Revista Científica (Journal Article)
Sitemtc-m21d.sid.inpe.br
Código do Detentorisadg {BR SPINPE} ibi 8JMKD3MGPCW/3DT298S
Identificador8JMKD3MGP3W34T/4A3ALEH
Repositóriosid.inpe.br/mtc-m21d/2023/10.17.19.10   (acesso restrito)
Última Atualização2023:10.17.19.10.39 (UTC) simone
Repositório de Metadadossid.inpe.br/mtc-m21d/2023/10.17.19.10.39
Última Atualização dos Metadados2024:01.02.17.16.49 (UTC) administrator
DOI10.3390/ai4030032
ISSN2673-2688
Chave de CitaçãoSantiagoJúnior:2023:EvDeLe
TítuloEvaluating Deep Learning Techniques for Blind Image Super-Resolution within a High-Scale Multi-Domain Perspective
Ano2023
MêsSept.
Data de Acesso01 jun. 2025
Tipo de Trabalhojournal article
Tipo SecundárioPRE PI
Número de Arquivos1
Tamanho4319 KiB
2. Contextualização
AutorSantiago Júnior, Valdivino Alexandre de
Identificador de Curriculo8JMKD3MGP5W/3C9JJB5
ORCID0000-0002-4277-021X
GrupoCOPDT-CGIP-INPE-MCTI-GOV-BR
AfiliaçãoInstituto Nacional de Pesquisas Espaciais (INPE)
Endereço de e-Mail do Autorvaldivino.santiago@inpe.br
RevistaAI
Volume4
Número3
Páginas598-619
Histórico (UTC)2023-10-17 19:11:04 :: simone -> administrator :: 2023
2024-01-02 17:16:49 :: administrator -> simone :: 2023
3. Conteúdo e estrutura
É a matriz ou uma cópia?é a matriz
Estágio do Conteúdoconcluido
Transferível1
Tipo do ConteúdoExternal Contribution
Tipo de Versãopublisher
Palavras-Chaveimage super-resolution
artificial intelligence
deep learning
controlled experiment
multiple domains
ResumoDespite several solutions and experiments have been conducted recently addressing image super-resolution (SR), boosted by deep learning (DL), they do not usually design evaluations with high scaling factors. Moreover, the datasets are generally benchmarks which do not truly encompass significant diversity of domains to proper evaluate the techniques. It is also interesting to remark that blind SR is attractive for real-world scenarios since it is based on the idea that the degradation process is unknown, and, hence, techniques in this context rely basically on low-resolution (LR) images. In this article, we present a high-scale (8×) experiment which evaluates five recent DL techniques tailored for blind image SR: Adaptive Pseudo Augmentation (APA), Blind Image SR with Spatially Variant Degradations (BlindSR), Deep Alternating Network (DAN), FastGAN, and Mixture of Experts Super-Resolution (MoESR). We consider 14 datasets from five different broader domains (Aerial, Fauna, Flora, Medical, and Satellite), and another remark is that some of the DL approaches were designed for single-image SR but others not. Based on two no-reference metrics, NIQE and the transformer-based MANIQA score, MoESR can be regarded as the best solution although the perceptual quality of the created high-resolution (HR) images of all the techniques still needs to improve.
ÁreaCOMP
Arranjourlib.net > CGIP > Evaluating Deep Learning...
Conteúdo da Pasta docacessar
Conteúdo da Pasta sourcenão têm arquivos
Conteúdo da Pasta agreement
agreement.html 17/10/2023 16:10 1.0 KiB 
4. Condições de acesso e uso
Idiomaen
Arquivo Alvoai-04-00032.pdf
Grupo de Usuáriossimone
Grupo de Leitoresadministrator
simone
Visibilidadeshown
Permissão de Leituradeny from all and allow from 150.163
Permissão de Atualizaçãonão transferida
5. Fontes relacionadas
Unidades Imediatamente Superiores8JMKD3MGPCW/46KUES5
Lista de Itens Citandosid.inpe.br/bibdigital/2022/04.03.23.11 8
sid.inpe.br/mtc-m21/2012/07.13.15.01.24 2
DivulgaçãoWEBSCI; PORTALCAPES.
Acervo Hospedeirourlib.net/www/2021/06.04.03.40
6. Notas
Campos Vaziosalternatejournal archivingpolicy archivist callnumber copyholder copyright creatorhistory descriptionlevel e-mailaddress format isbn label lineage mark mirrorrepository nextedition notes parameterlist parentrepositories previousedition previouslowerunit progress project rightsholder schedulinginformation secondarydate secondarykey secondarymark session shorttitle sponsor subject tertiarymark tertiarytype url
7. Controle da descrição
e-Mail (login)simone
atualizar 


Fechar