Title :
An overview of the SPHINX speech recognition system
Author :
Lee, Kai-Fu ; Hon, Hsiao-Wuen ; Reddy, Raj
Author_Institution :
Sch. of Comput. Sci., Carnegie Mellon Univ., Pittsburgh, PA, USA
fDate :
1/1/1990 12:00:00 AM
Abstract :
A description is given of SPHINX, a system that demonstrates the feasibility of accurate, large-vocabulary, speaker-independent, continuous speech recognition. SPHINX is based on discrete hidden Markov models (HMMs) with LPC- (linear-predictive-coding) derived parameters. To provide speaker independence, knowledge was added to these HMMs in several ways: multiple codebooks of fixed-width parameters, and an enhanced recognizer with carefully designed models and word-duration modeling. To deal with coarticulation in continuous speech, yet still adequately represent a large vocabulary, two new subword speech units are introduced: function-word-dependent phone models and generalized triphone models. With grammars of perplexity 997, 60, and 20, SPHINX attained word accuracies of 71, 94, and 96%, respectively, on a 997-word task
Keywords :
Markov processes; speech recognition; HMM; LPC; SPHINX speech recognition system; coarticulation; continuous speech recognition; fixed-width parameters; function-word-dependent phone models; generalized triphone models; grammars; hidden Markov models; large-vocabulary; linear-predictive-coding; multiple codebooks; speaker-independent; subword speech units; word accuracies; word-duration modeling; Automatic speech recognition; Continuous wavelet transforms; Degradation; Error analysis; Hidden Markov models; Marine vehicles; Mars; Speech recognition; System testing; Vocabulary;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
