An Intelligent Hearing Aid System Based on Real-Time Signal Processing

Abstract

A hearing aid, worn in or around a wearer's ear, is an electroacoustic device to block noise and enhance target sound signal for the wearer. However, there is no method in traditional hearing aids that can automatically separate noise from a received signal mixture. Especially when the backing round is complex, amplified sound by a traditional hearing aid is still a multi-source mixture, and its wearer could not extract the target sound as expected. In this paper, a smart hearing aid system is proposed, consisting of an acoustic property modeling module, an acoustic source separation module and a target source recognition module. The proposed solution can automatically separate an acoustic source and thus enhance the target speaker's sound and filter the other sounds. Experiments with three sources has been conducted and the results show the feasibility and validity of our intelligent hearing aid.