Evaluating Automatic Speech Recognition as a Conversational Aid for People with Hearing Loss
By Harkins, Judith E., PhD; Haravon, Anita B.; Tucker, Paula E.; Eum, Julianne; Fisher, Laura; RESNA 2001: Annual Conference Proceedings, Vol. 21, No. 1, pp. 106-108Publication Date: June 2001
Paper presents partial results of a project examining communication dynamics, system performance, and user satisfaction with automatic speech recognition (ASR) used in live conversation between hearing people and people who are deaf or hard of hearing. The main objectives of the study were to evaluate how accurately ASR transcribes sustained, unscripted conversational speech, and to understand users' attitudes toward this method when compared with other face-to-face methods and technologies. Six participants were used in the study, all being adults who had social or professional familiarity with people who are deaf and hard of hearing, and knowledge of the Microsoft Windows operating system, but no prior experience with ASR. The participants received two hours of training and practice with Dragon NaturallySpeaking 4.01. These participants were paired with six others with severe-profound hearing loss. The participants were evaluated based on conversation accuracy between hearing participants and participants who were hearing-impaired while using ASR. Accuracy ranged from 60 to 97 percent in every conversation, although every user reported that they preferred communication without the use of technology.
Assistive Products Discussed: DRAGON NATURALLYSPEAKING
Published by: Rehabilitation Engineering & Assistive Technology Society of North America (RESNA) (Website:http://www.resna.org)
ISBN: 0-932101-43-7
This publication is included in the library of the National Rehabilitation Information Center (NARIC), accession number O14163
