Automated Speech Recognition for Captioned Telephone Conversations

Authors

Jeff Adams CEO, Cobalt Speech and Language, IncFollow
Kenneth Basye PhD, Clark UniversityFollow
Alok Parlikar PhD, Cobalt Speech and LanguageFollow
Andrew Fletcher PhD, Cobalt Speech and LanguageFollow
Jangwon Kim PhD, Canary Speech, LLCFollow

Document Type

Article

Abstract

Internet Protocol Captioned Telephone Service is a service for people with hearing loss, allowing them to communicate effectively by having a human Communications Assistant transcribe the call and equipment that displays the transcription in near real time. The current state of the art for ASR is considered with regard to automating such service. Recent results on standard tests are examined and appropriate metrics for ASR performance in captioning are discussed. Possible paths for developing fully-automated telephone captioning services are examined and the effort involved is evaluated.

Publication Date

11-3-2017

Keywords

speech recognition captioned telephone service transcription

Repository Citation

Adams, Jeff CEO; Basye, Kenneth PhD; Parlikar, Alok PhD; Fletcher, Andrew PhD; and Kim, Jangwon PhD, "Automated Speech Recognition for Captioned Telephone Conversations" (2017). Faculty Works. 26.
https://commons.clarku.edu/facultyworks/26