The use of acoustic contextual information in HMM-based speech recognition

Author

Choi, In-Jeong ; Lee, Soo-Young

Author_Institution

Dept. of Electr. Eng., Korea Adv. Inst. of Sci. & Technol., Taejon, South Korea

Volume

5

Issue

5

fYear

1998

fDate

5/1/1998 12:00:00 AM

Firstpage

108

Lastpage

110

Abstract

A novel method is proposed to incorporate acoustic contextual information into speech recognition systems based on the hidden Markov model (HMM). Frame correlation exponents and transition costs are introduced to measure the effects of contextual information and modify maximum likelihood solutions in standard HMMs. The contextual information parameters reflect both time correlation among feature vectors and boundary effects between HMM states. Significant reduction of error rates is achieved for a continuous speech recognition task.

Keywords

acoustic correlation; hidden Markov models; speech recognition; HMM-based speech recognition; acoustic contextual information; boundary effects; continuous speech recognition task; error rate reduction; feature vectors; frame correlation exponents; hidden Markov model; maximum likelihood solutions; time correlation; transition costs; Acoustic measurements; Computational complexity; Costs; Error analysis; Hidden Markov models; Maximum likelihood estimation; Measurement standards; Parameter estimation; Speech recognition; Uncertainty;

fLanguage

English

Journal_Title

Signal Processing Letters, IEEE

Publisher

ieee

ISSN

1070-9908

Type

jour

DOI

10.1109/97.668943

Filename

668943