Please use this identifier to cite or link to this item: https://repository.cihe.edu.hk/jspui/handle/cihe/1239
DC FieldValueLanguage
dc.contributor.authorSiu, Wan Chien_US
dc.contributor.authorLiu, Zhisong-
dc.contributor.otherChan, Y.-L.-
dc.date.accessioned2021-08-11T05:54:32Z-
dc.date.available2021-08-11T05:54:32Z-
dc.date.issued2021-
dc.identifier.urihttps://repository.cihe.edu.hk/jspui/handle/cihe/1239-
dc.description.abstractThere is a great leap in objective accuracy on image super-resolution, which recently brings a new challenge on image super-resolution with larger up-scaling (e.g. 4× ) using pixel based distortion for measurement. This causes over-smooth effect which cannot grasp well the perceptual similarity. The advent of generative adversarial networks makes it possible super-resolve a low-resolution image to generate photo-realistic images sharing distribution with the high-resolution images. However, generative networks suffer from problems of mode-collapse and unrealistic sample generation. We propose to perform Image Super-Resolution via Variational AutoEncoders (SR-VAE) learning according to the conditional distribution of the high-resolution images induced by the low-resolution images. Given that the Conditional Variational Autoencoders tend to generate blur images, we add the conditional sampling mechanism to narrow down the latent subspace for reconstruction. To evaluate the model generalization, we use KL loss to measure the divergence between latent vectors and standard Gaussian distribution. Eventually, in order to balance the trade-off between super-resolution distortion and perception, not only that we use pixel based loss, we also use the modified deep feature loss between SR and HR images to estimate the reconstruction. In experiments, we evaluated a large number of datasets to make comparison with other state-of-the-art super-resolution approaches. Results on both objective and subjective measurements show that our proposed SR-VAE can achieve good photo-realistic perceptual quality closer to the natural image manifold while maintain low distortion.en_US
dc.language.isoenen_US
dc.publisherIEEEen_US
dc.relation.ispartofIEEE Transactions on Circuits and Systems for Video Technologyen_US
dc.titlePhoto-realistic image super-resolution via variational autoencodersen_US
dc.typejournal articleen_US
dc.identifier.doi10.1109/TCSVT.2020.3003832-
dc.contributor.affiliationSchool of Computing and Information Sciencesen_US
dc.relation.issn1558-2205en_US
dc.description.volume31en_US
dc.description.issue4en_US
dc.description.startpage1351en_US
dc.description.endpage1365en_US
dc.cihe.affiliatedNo-
item.openairecristypehttp://purl.org/coar/resource_type/c_6501-
item.cerifentitytypePublications-
item.openairetypejournal article-
item.fulltextNo Fulltext-
item.grantfulltextnone-
item.languageiso639-1en-
crisitem.author.deptSchool of Computing and Information Sciences-
crisitem.author.deptSchool of Computing and Information Sciences-
crisitem.author.orcid0000-0001-8280-0367-
crisitem.author.orcid0000-0003-4507-3097-
Appears in Collections:CIS Publication
SFX Query Show simple item record

Google ScholarTM

Check

Altmetric

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.