Publicación:
Generative Adversarial Networks for text-to-face synthesis & generation: A quantitative–qualitative analysis of Natural Language Processing encoders for Spanish

Vista sencilla de ítem
dc.contributor.authorYauri Lozano, Eduardo
dc.contributor.authorOrozco Barbosa, Luis
dc.contributor.authorGarcía Castro, Raúl
dc.contributor.authorCastillo Cara, José Manuel
dc.date.accessioned2024-05-20T11:38:08Z
dc.date.available2024-05-20T11:38:08Z
dc.date.issued2024-01
dc.description.abstractIn recent years, the development of Natural Language Processing (NLP) text-to-face encoders and Generative Adversarial Networks (GANs) has enabled the synthesis and generation of facial images from textual description. However, most encoders have been developed for the English language. This work presents the first study of three text-to-face encoders, namely, the RoBERTa pre-trained model and the Sent2Vec and RoBERTa models, trained with the CelebA dataset in Spanish. It then introduces customised and fine-tuned conditional Deep Convolutional Generative Adversarial Networks (cDCGANs) trained with the CelebA dataset for text-to-face generation in Spanish. To validate the results obtained, a qualitative evaluation was carried out with a visual analysis and a quantitative evaluation based on the IS, FID and LPIPS metrics. Our findings show promising results with respect to the literature, improving the numerical metrics of FID and LPIPS by 5% and 37%, respectively. Our results also show, through a quantitative–qualitative comparison of the cDCGAN training epochs, that the IS metric is not a reliable objective metric to be considered in the evaluation of similar worksen
dc.description.versionversión publicada
dc.identifier.doihttps://doi.org/10.1016/j.ipm.2024.103667
dc.identifier.issn0306-4573 eISSN 1873-5371
dc.identifier.urihttps://hdl.handle.net/20.500.14468/12342
dc.journal.issue3
dc.journal.titleInformation Processing and Management
dc.journal.volume61
dc.language.isoes
dc.publisherElsevier
dc.relation.centerE.T.S. de Ingeniería Informática
dc.relation.departmentInformática y Automática
dc.rightsinfo:eu-repo/semantics/openAccess
dc.rights.urihttps://creativecommons.org/licenses/by-nc-nd/4.0/deed.es
dc.subject.keywordsmage synthesis
dc.subject.keywordsCelebA dataset
dc.subject.keywordsRoBERTa transformer
dc.subject.keywordsSpanish
dc.subject.keywordscDCGAN
dc.subject.keywordsText-to-face generation
dc.subject.keywordsText-to-image synthesis
dc.titleGenerative Adversarial Networks for text-to-face synthesis & generation: A quantitative–qualitative analysis of Natural Language Processing encoders for Spanishes
dc.typejournal articleen
dc.typeartículoes
dspace.entity.typePublication
relation.isAuthorOfPublicationc0e39bd2-c0d8-4743-953d-488baf6b977e
relation.isAuthorOfPublication.latestForDiscoveryc0e39bd2-c0d8-4743-953d-488baf6b977e
Archivos
Bloque original
Mostrando 1 - 1 de 1
Miniatura
Nombre:
Castillo_Cara_Jose_Manuel_GANs.pdf
Tamaño:
1.74 MB
Formato:
Adobe Portable Document Format
Descargar
Colecciones
Artículos y papers