Principles of electronic speech processing with applications for people with disabilities

Abstract

In the first part of this paper the principles and the state of the art of speech processing, and especially speech synthesis and recognition, are explained. Then, a speech-based human-computer dialogue system is discussed. The next section gives a brief overview of the available recommendations, guidelines and standards that are directly related with the application of speech technologies. The last part of the paper is dedicated to applications of speech technology for the disabled. The main focus is on blind and partially sighted people and those with hearing loss. Concerning the blind, many multilingual text-to-speech synthesis systems exist, and some polyglot ones, that can convert printed and electronic documents to audio, but further research is needed for structured text, tables and above all graphics to be efficiently transformed into speech. For the deaf persons, there are still big challenges in the development of adequate communication aids. Although a high-speed transformation of speech into text is possible with state-of-the-art speech recognizers (and thus a quasi real-time information transfer from a hearing to a deaf person), the automatic gesture recognition, needed for the reverse transfer, is still in research state. Other applications discussed in this paper include speech-based cursor control for those with physical disabilities, transformation of dysarthric speech into intelligible speech, voice output communication aids for the language impaired and those without speech, and accessibility options for public terminals and Automated Teller Machines through the incorporation of speech technologies. The paper concludes with an outlook and recommendations for research areas that need further study.