Work is proposed to develop a communication interface for facilitating telephone communication between individuals with severe speech and/or hearing communication impairments (text users) and individuals with normal speech and hearing function (voice users). The interface would be a computer-based augmentation of the telephone at the location of the text user; it would require no special equipment at the voice user's end. The interface would include automatic speech recognition (ASR) and/or automatic text-to-speech synthesis (TTS). The ASR would transform the voice user's speech signal received over the phone to visually displayed text; the TTS system would transform text output produced by the text user to speech for transmission over the phone to the voice user. Research in Phase I will test the feasibility of this communication interface. A key component in this work is a state-of-the-art, large-vocabulary, telephony- based, continuous speech recognizer. With this ASR, the proposed work is aimed at developing an interface that will provide for easy and rapid recovery from errors, making fluent communication between the voice user and the text user possible. To the extent that the envisioned interface can be successfully realized, it will enhance communication access and privacy for communicatively-impaired persons and their relatives, friends, and associates.
The market for the proposed interface system would be the several million people with severe hearing impairment or speech impairment who rely on the telephone relay system.