VecGAN: Image-to-Image translation with interpretable latent directions

Simple item page
buir.contributor.authorDalva, Yusuf
buir.contributor.authorDundar, Aysegul
buir.contributor.authorAltındiş, Said Fahri
buir.contributor.orcidDalva, Yusuf|0000-0002-8402-8291
dc.citation.epage169en_US
dc.citation.spage153en_US
dc.citation.volumeNumber13676en_US
dc.contributor.authorDalva, Yusuf
dc.contributor.authorDundar, Aysegul
dc.contributor.authorAltındiş, Said Fahri
dc.date.accessioned2023-02-16T08:52:00Z
dc.date.available2023-02-16T08:52:00Z
dc.date.issued2022-10-21
dc.departmentDepartment of Computer Engineeringen_US
dc.description.abstractWe propose VecGAN, an image-to-image translation framework for facial attribute editing with interpretable latent directions. Facial attribute editing task faces the challenges of precise attribute editing with controllable strength and preservation of the other attributes of an image. For this goal, we design the attribute editing by latent space factorization and for each attribute, we learn a linear direction that is orthogonal to the others. The other component is the controllable strength of the change, a scalar value. In our framework, this scalar can be either sampled or encoded from a reference image by projection. Our work is inspired by the latent space factorization works of fixed pretrained GANs. However, while those models cannot be trained end-to-end and struggle to edit encoded images precisely, VecGAN is end-to-end trained for image translation task and successful at editing an attribute while preserving the others. Our extensive experiments show that VecGAN achieves significant improvements over state-of-the-arts for both local and global edits.en_US
dc.description.provenanceSubmitted by Aleyna Demirkıran (demirkiranaleyna99@gmail.com) on 2023-02-16T08:52:00Z No. of bitstreams: 1 VecGAN_Image_to_Image_Translation_with_Interpretable_Latent_Directions.pdf: 4027888 bytes, checksum: 827bf1fef5ec2bafa0b68e9a30efb404 (MD5)en
dc.description.provenanceMade available in DSpace on 2023-02-16T08:52:00Z (GMT). No. of bitstreams: 1 VecGAN_Image_to_Image_Translation_with_Interpretable_Latent_Directions.pdf: 4027888 bytes, checksum: 827bf1fef5ec2bafa0b68e9a30efb404 (MD5) Previous issue date: 2022-10-21en
dc.identifier.doi10.1007/978-3-031-19787-1_9en_US
dc.identifier.issn03029743
dc.identifier.urihttp://hdl.handle.net/11693/111419
dc.language.isoEnglishen_US
dc.relation.isversionofhttps://www.doi.org/10.1007/978-3-031-19787-1_9en_US
dc.source.titleComputer Vision – ECCV 2022en_US
dc.subjectImage translationen_US
dc.subjectGenerative adversarial networksen_US
dc.subjectLatent space manipulationen_US
dc.subjectFace attribute editingen_US
dc.titleVecGAN: Image-to-Image translation with interpretable latent directionsen_US
dc.typeArticleen_US

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
VecGAN_Image_to_Image_Translation_with_Interpretable_Latent_Directions.pdf
Size:
3.84 MB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.69 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Department of Computer Engineering