Um serviço de gestão de dicionário para a suíte VLibras

Abstract

Deaf people face many barriers to be granted access to information. Their inclusion, however, is permeated by the knowledge that their first language is the sign language, given that oral dialects are considered their second language. Because of this, they might have difficulties reading and writing in their country’s official language, creating a series of information access and communication problems. In addition to this, information and communication technologies (ICTs) hardly ever take these users specific needs into account when designed. For instance, TV support for sign language is usually limited to a small window on the corner of the screen with a human interpreter, generating high operational costs. Besides, manual translation to sign language is not usually effective on environments such as the Web, with dynamic, rapidly changing content. To address these issues, there are a few automatic translation tools available on the market, but they are generally closed and don’t offer public APIs for integration with other services. This objective of this work is to develop a service called DaaS (Dictionary as a Service) that will provide control and management of the signs in the automatic translation tool suite VLIBRAS, turning it into an open dictionary for the Brazilian Sign Language (LIBRAS). Thus, it is intended to facilitate the creation of services containing dictionaries of LIBRAS, the integration of automatic translation tools to LIBRAS, among others. Using this service, users will be able to create their own LIBRAS dictionary through videos, animations with a 3D avatar or metadata corresponding to the signs; make version control, copies and searches on this dictionary, among others. To develop this service, the necessities of the users needs were analyzed, functional and non-functional requirements documented and an architecture based on these requirements was created. The service based on these architecture and requirements was developed using the JavaScript programming language with Node.js and MongoDB. DaaS integration with VLIBRAS was validated through a pretest using the update control of the VLibras applications sign languages dictionary.