Augmented reality speech recognition for the hearing impaired

Our paper presents a novel Speech Recognition Augmented Reality tool for hearing impaired people, based on a case study of a simple Domain Based Speech Recognition approach. Choosing the correct word out of samples with similar acoustic structure is an important challenge in Speech Recognition research. Our scheme improves transcription accuracy in local vocabulary environments, such as in meetings, lectures etc. We developed a tool that combines our approach in an Augmented Reality environment in order to create real-time “live subtitles” using our unique Speech Recognition process. This tool enables hearing impaired and deaf people to see real-time Augmented Reality subtitles while hearing a talk on a specific subject.