CLIPTraVeLGAN for Semantically Robust Unpaired Image Translation

Bodyanskiy, Y.; Ryabova, N.; Lavrynenko, R.

Публікація:
CLIPTraVeLGAN for Semantically Robust Unpaired Image Translation

dc.contributor.author	Bodyanskiy, Y.
dc.contributor.author	Ryabova, N.
dc.contributor.author	Lavrynenko, R.
dc.date.accessioned	2023-05-11T14:04:01Z
dc.date.available	2023-05-11T14:04:01Z
dc.date.issued	2023
dc.description	The work was performed as part of the state budget research project “Development of methods and algorithms for combined learning of deep neuro-neo-fuzzy systems under short training set conditions” (state registration number 0122U001701) of Artificial Intelligence Department of Kharkiv National University of Radio Electronics.
dc.description.abstract	In this paper a novel approach for semantically robust unpaired image translation is presented. CLIPTraVeLGAN replaces the Siamese network in TraVeLGAN with a contrastively pretrained language-image model (CLIP) with frozen weights. This approach significantly simplifies the model selection and training process of TraVeLGAN, making it more robust and easier to use.
dc.identifier.citation	Bodyanskiy Y. CLIPTraVeLGAN for Semantically Robust Unpaired Image Translation / Y. Bodyanskiy, N. Ryabova, R. Lavrynenko // Computational Linguistics and Intelligent Systems (COLINS-2023) : Proc. 7th Int. Conf., April 20–21, 2023. – Kharkiv, 2023. – Volume I (Machine Learning Workshop). – pp.1-12.
dc.identifier.other	urn:nbn:de:0074-3387-8
dc.identifier.uri	https://openarchive.nure.ua/handle/document/22867
dc.language.iso	en
dc.subject	Image-to-image translation, GAN, CLIP, Transfer knowledge
dc.title	CLIPTraVeLGAN for Semantically Robust Unpaired Image Translation
dc.type	Conference proceedings
dspace.entity.type	Publication