Computer Vision-based Bangladeshi Sign Language Recognition System

Abstract

Sign language is a specific area of human gesture communication and a full-edged complex language that is used by various deaf communities. In Bangladesh, there are many deaf and dumb people. It becomes very difficult to communicate with them for the people who are unable to understand the Sign Language. In this case, an interpreter can help a lot. So it is desirable to make computer to understand the Bangladeshi sign language that can serve as an interpreter. In this paper, a Computer Vision-based Bangladeshi Sign Language Recognition System (BdSL) has been proposed. In this system, separate PCA (Principal Component Analysis) is used for Bengali Vowels and Bengali Numbers recognition. The system is tested for 6 Bengali Vowels and 10 Bengali Numbers.