Real-Time Application for Deaf and Dumb: Bidirectional Communication Using Learning Methods

Abstract

AbstractIn this world, we have more than 466 million deaf dumb peoples. Normal peoples are communicating with voice, likewise deaf and dumb people are communicating by the use of sign language. Generally, without knowing the language, one cannot talk to other people in the world, similarly without knowing the sign language normal people cannot talk to deaf and dumb people. Within themselves, deaf and dumb people will communicate using sign language, which normal people cannot understand, but still hearing is one way to understand the people, so expression and gesture are also a way to understand thoughts of deaf and dumb people. To find a solution for this problem, either normal person needs to learn the sign language or we need to create a platform which will convert the sign language to voice and voice to sign language. Though we have many different types platform, still it not efficient so we need to use the advanced method which recognizes the input much efficiently. Since we are using gesture and voice as an input, this will act as bidirectional way communication. The focus of this work is to create a bidirectional-based system to identify sign language gestures from the sign language images and converting that to voice output. Since we are developing the application to computer and android, it should have two-way communication so that both can act as input and output node. For efficient voice input and output, we are using android mobile.KeywordsGestureSign languageBidirectionalDeaf and dumbTwo-way communication systemComputer visionSIFT algorithm