Falibras Messenger: Solution to the Accessibility of the Deafs in Telegram Web with the Aid of Volunteer Computer Grid

Abstract

In Brazil there are about 9.6 million people who have some kind of hearing impairment. The Brazilian deaf own mother tongue the Brazilian Sign Language - Libras. Because it is a visual language, a deaf person has great difficulty in social interaction, particularly with hearing people. One way to work around this limitation is by supporting tools, such as Falibras Messenger, presented in this paper. It is an integrated Portuguese-Libras translator to Telegram Web application that aims to facilitate communication between deaf and hearing people. However, to enable the use of Falibras Messenger, it was necessary to refactor its architecture to adapt it to new scalability requirements, given the increasing demand of requests for instant messaging applications. In this work, the development of a new distributed translation architecture based on volunteer grid computing and a new tool for communication of deaf and hearing on the Web Telegram is addressed. The developed architecture was evaluated in a heterogeneous experimental environment presenting gains in scalability and increased performance of the median time translation.