Animador: um serviço de geração semiautomática de sinais em libras

Abstract

The deaf suffer several problems in regard to access of information, especially those information circulating in Information and Communication Technologies (ICTs). These technologies are usually designed for people who understand the spoken language, whereas for the deaf, the oral language is considered something like a "second language". On the other hand, companies responsible for making such information available feel great difficulty in making accessible the content provided by them, especially since these contents are generally dynamic, and the production of windows in sign languages (eg, Brazilian Sign Language - Libras) has usually a high operating cost. To minimize these problems, an alternative is the use of automatic translators to sign languages. With these automatic translators, it can generate digital content dynamically, with low cost, and according to some studies, with a reasonable standard of intelligibility. These solutions, however, generally require a database (dictionary LBS) containing thousands of videos or animations represented by an animated avatar. The process of building these databases, however, is a arduous process, and depends on a large and specialized team of modelers, animators and 3D designers, as well as experts on sign language and deaf users. To minimize these problems, the purpose of this work is to develop a solution for semi-automatic generation of these animations (movies). The idea is that these animations can be created by users from the parameter setting (phonemes) of the Brazilian Sign Language (Libras), reducing the time and cost required to produce this kind of database.