DMCA
An overview of the SPHINX speech recognition system (1990)
| Citations: | 163 - 9 self |
BibTeX
@MISC{Lee90anoverview,
author = {Kai-fu Lee and Hsiao-wuen Hon and A N D Raj Reddy and Fellow Ieee},
title = {An overview of the SPHINX speech recognition system},
year = {1990}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract-Speaker independence, continuous speech, and large vo-cabularies pose three of the greatest challenges in automatic speech recognition. Previously, accurate speech recognizers avoided dealing simultaneously with all three problems. This paper describes SPHINX, a system that demonstrates the feasibility of accurate, large-vocabu-lary speaker-independent, continuous speech recognition. SPHINX is based on discrete hidden Markov models (HMM’s) with LPC-derived parameters. To provide speaker independence, we added knowledge to these HMM’s in several ways: multiple codebooks of fixed-width parameters, and an enhanced recogniar with carefully de-signed models and word duration modeling. To deal with coarticula-tion in continuous speech, yet still adequately represent a large vocab-ulary, we introduce two new subword speech units-function-word-dependent phone models and generaliied triphone models. With gram-mars of perplexity 997, 60, and 20, SPHINX attained word accuracies
Keyphrases
sphinx speech recognition system continuous speech de-signed model abstract-speaker independence accurate speech recognizers lpc-derived parameter multiple codebooks enhanced recogniar fixed-width parameter large-vocabu-lary speaker-independent large vo-cabularies several way automatic speech recognition triphone model continuous speech recognition word accuracy speaker independence word duration discrete hidden markov model
