Study for implementing an automatic translation sofware for libras

• Autor(a) principal: Sigrist, Vanina Carrara
• Data de Publicação: 2021
• Outros Autores: Siqueira, Anna Haydée de Lima, Cunha, Danubia Pinheiro
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Processando o Saber (Online)
• Texto Completo: https://fatecpg.edu.br/revista/index.php/ps/article/view/163
• topic: Acessibilidade; Inteligência Artificial; Reconhecimento Óptico de Caracteres; Rede neural artificial; Língua brasileira de sinais (LIBRAS); Accessibility; Artificial intelligence; Optical character recognition (OCR); Machine translation; Artificial neural network; Brazilian Sign Language

This article aims to identify how artificial intelligence together with optical character recognition (OCR) can help in the translation of documents and books written in Portuguese into the Brazilian Sign Language. Through scientific articles and theses published on the two technologies, we analyzed the possibility of applying both in the translation process, automating it and making it more accessible to the people who are Deaf or Hard of Hearing (DHH), whether they are literate in Portuguese or not. For this article, we performed theoretical research and a bibliographic review to present the application of these technologies in the proposed scenario. Finally, based on what has been evidenced, we concluded the feasibility of an application that uses separate modules of OCR and a neural network interconnected to perform automatic translation.