Video-based turkish sign language recognition systems

Abstract

Automatic recognition of sign language will be extremely beneficial for hearing impaired people, by allowing them to easily communicate with the outside world. Sign language is a non-verbal and visual language, which evolved detached from the spoken language and it is based on the simultaneous contribution of both temporal and spatial configurations. Even if it is spread all over the world, sign language is not a universal language and every country needs to build its own automatic system. This paper challenges the issue of the development of automatic video-based Turkish Sign Language (TSL) recognition systems; it presents two computer vision-based automatic systems for TSL recognition and compares their performances. Since we are using a newly released TSL database, this paper has also the side-effect to divulge the database to the research community. © 2016 Taylor & Francis Group, London.