An Automatic Emergency Signal Recognition System for the Hearing Impaired

Abstract

Communication systems based on man-machine interaction play an increasingly important role in everyday life. Whereas these systems make life easier for many people, for the hearing impaired they are indispensable. Automatic speech recognition systems in particular, thanks to the simultaneous display of text, reduce the gap separating the deaf, or the hearing impaired in general, from the world of the hearing. The paper presents an automatic aid system for the hearing impaired while behind the wheel of a car. It proposes an algorithm, based on typical speech recognition techniques, for the identification and automatic signaling on a display of emergency signals such as sirens sounded by ambulances, fire engines or any other emergency service vehicle. Tests performed show that the system guarantees an accuracy level of over 99% with a response time of less than 400 ms