HamNoSys based Indian Sign Language Generation System

Abstract

Sign language (SL) is the most natural way of expression of thoughts and feelings for the deaf community. Indian SL is a gestural language which provides information by using hands and other body parts. In India, there are approximately over 24 million people who are deaf or hard of hearing, out of which only 5% attend schools. The reason for the lack of language and literacy skills among hearing impaired people is that there are very few sources in India that can provide education to them. There is a need of automated systems that can overcome this problem, thus improving the communication skills among deaf communities. In this thesis, a system for creation of HamNoSys based ISL dictionary has been proposed. The system has a virtual keyboard for entering HamNoSys of a given sign. The sign gloss and the HamNoSys of a given word are stored in ISL dictionary. There are various signs in Indian Sign Language that make use of non-manual sign gestures. A system for selecting non-manual components such as eye-gaze, mouth gestures, face expressions and body postures for a given sign is also developed and embedded in the proposed system. The sign and its HamNoSys that are stored in the dictionary can be retrieved from it by the users. The system has a translator that converts the retrieved HamNoSys to SiGML. The generated SiGML file is then given as an input to the signing avatar which animates the signs. Thus in this way the proposed system can be used for creating Indian sign language dictionary as well as for animating the signs also.