An Effective Approach to Communicate with the Deaf and Mute People by Recognizing Characters of One-hand Bangla Sign Language Using Convolutional Neural-Network

Abstract

Sign language is one of the best communication medium for Deaf and Mute people who can not speak to others or hear from others. In most of the cases, the relatives or family members of Deaf and Mute person face difficulties to express their opinion and to communicate with them. Therefore, it is not easy to learn sign language for communicating with the Deaf and Mute individuals. Thus, an interpreter is essential to interact with deaf and mute people who can interpret hand gestures to characters and characters to hand gestures. However, the appointment of an expert interpreter for most of the families having deaf and mute members in a low incoming country like Bangladesh. Considering these issues, we propose a system in which all members in a family where deaf and mute family members exists can communicate easily and efficiently. In our proposed system, we have used convolutional neural-network to recognize hand gestures and classify the characters and vice versa. This recognition is so swift that an instant communication system can be developed in which continuous conversion becomes effortless.