A Computer-Based Application for Speech Recognition in Multi-Speaker Environment to Assist Hearing Impaired People

Abstract

In this paper, a computer-based application has been proposed that facilitates speech recognition in a multi-speaker environment. It identifies the speaker, recognizes the speech and displays the text output obtained from speech to text conversion along with the identified speaker’s name tag on the computer screen. The proposed system provides three functionalities: i) Speech Recognition ii) Speaker Identification iii) Many to one transmission-reception of converted speech to text using the client-server model and multi-threading concept. The system consists of transmission and reception capable devices like computers, a computer application that connects various devices using WiFi technology and microphone to obtain speech input from the users. All the functionalities are implemented as a computer application with a user-friendly graphical user interface(GUI). The application has two scenarios: i) Near and ii) Far. In the case when the speakers are in the range of the user’s microphone, the user opts for near scenario else user opts for the far scenario. The proposed system is intended mainly for people suffering from hearing loss. Although hearing aid technology is greatly improved, it still utilizes an impaired person’s inner, outer and middle ear including hearing nerve to support natural hearing.