Indexação automática de imagens na web: tendências e desafios no contexto Deep Learning

• Autor(a) principal: Gracioso, Luciana de Souza
• Data de Publicação: 2018
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista Ibero-americana de Ciência da Informação
• Texto Completo: https://periodicos.unb.br/index.php/RICI/article/view/8342
• topic: deep learning; indexação de imagens; machine learning; recuperação de imagens na web; deep learning; indización de imágenes; machine learning; recuperación de imágens en la web; deep learning; image retrieval on the web; indexing of imagens; machine learning

The objective of this study is to investigate the extent to which research in Information Science (IS) has approximated those techniques of the Deep Learning, being related to representation, description and retrieval of images on the Web, and thus, to assess the value of these researches when applied to IS methods. From an integrative review of national and international literature contextualized in the IS, the recovered documents were analyzed according to the criteria of the integrative review, identifying a set of operations that could be attached in the methodologies of representation and description of images developed and consolidated in the field of IS. It is concluded that there is still a gap in research of IS area both at national and international level on Deep Learning and that resources of this new programming structure can be approximated to the methods already validated by the area.