Challenges in Developing a Voice Input Voice Output Communication Aid for People With Severe Dysarthria
By Hawley, Mark; Cunningham, Stuart; Cardinaux, Fabien; Coy, Andre; O'Neill, Peter, Seghal, Siddharth; Enderby, Pam; Challenges for Assistive Technology: AAATE 2007 Conference Proceedings, pp. 363-367Publication Date: 2007
Paper describes the development of a voice-input voice-output communication aid (VIVOCA) for people with disordered or unintelligible speech. The device is in development for people with disabilities such as traumatic brain injury, cerebral palsy and progressive neurological conditions such as multiple sclerosis, which affect speech articulation. The VIVOCA is intended to recognize and interpret the user’s disordered speech and speak out an equivalent message in clear synthesized speech. The system comprises an input Bluetooth headset, an automatic speech recognition (ASR) component, a translation algorithm, a speech synthesizer, and an output speaker. After consultation with 8 VOCA users, a pocket computer (PDA) was chosen as the hardware platform. Methods for translating input speech into output speech explored include direct word-to-word translation, spelling, symbolic grids, and the building of phrases from word combinations. Implementation of the high-complexity algorithm of ASR in a low-computation, low-memory PDA was resolved using fixed-point calculations, resulting in a short sequence of input words that can be recognized in less than a second. The implemented algorithm also allows the user to get visual and audio feedback from the last pronounced word and to delete it if necessary. A VIVOCA prototype is being tried by users in order to optimize the design prior to a larger trial.
Published by: IOS Press (Website:http://www.iospress.nl)
Association for the Advancement of Assistive Technology in Europe (AAATE) (Web Site: http://www.aaate.net )
Link to text: http://www.booksonline.iospress.nl/Content/View.aspx?piid=7381
ISBN: 978-1-58603-791-8
