Robust methods for using context-dependent features and models in a continuous speech recognizer

Author

Bahl, L.R. ; de Souza, P.V. ; Gopalakrishnan, P.S. ; Nahamoo, D. ; Picheny, M.A.

Author_Institution

IBM Thomas J. Watson Res. Center, Yorktown Heights, NY, USA

Volume

i

fYear

1994

fDate

19-22 Apr 1994

Abstract

In this paper we describe the method we use to derive acoustic features that reflect some of the dynamics of frame-based parameter vectors. Models for such observations must be context dependent. Such models were outlined in an earlier paper. Here we describe a method for using these models in a recognition system. The method is more robust than using continuous parameter models in recognition. At the same time it does not suffer from the possible information loss in vector quantization based systems

Keywords

context-sensitive grammars; decoding; frame based representation; speech coding; speech recognition; vectors; acoustic features; context dependent models; context-dependent features; continuous parameter models; continuous speech recognizer; frame-based parameter vectors; information loss; models; recognition system; robust methods; vector quantization based systems; Context modeling; Data mining; Filter bank; Frequency; Hidden Markov models; Linear predictive coding; Robustness; Speech recognition; Training data; Vocabulary;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1994. ICASSP-94., 1994 IEEE International Conference on

Conference_Location

Adelaide, SA

ISSN

1520-6149

Print_ISBN

0-7803-1775-0

Type

conf

DOI

10.1109/ICASSP.1994.389238

Filename

389238