An improved VQ codebook design algorithm for HMM

Author

Koo, J.M. ; Lee, H.S. ; Un, C.K.

Author_Institution

DigiCom Inst. of Telematics, Seoul, South Korea

Volume

1

fYear

1992

fDate

23-26 Mar 1992

Firstpage

357

Abstract

The authors present a vector quantization (VQ) codebook design algorithm for recognition systems based on the phoneme-level hidden Markov model (HMM). In order to improve the discrimination ability of HMMs, the VQ codewords are extracted from the HMM state segments of the recognition units. An iterative VQ codebook design method based on the LVQ2 or MKM algorithm is proposed for the unified estimation of VQ codebooks and HMM parameters. Although the codebooks designed by the proposed algorithm showed larger quantization distortions for input speech spectrum than the conventional codebooks, they always improved the recognition accuracy of the recognition systems as compared to the conventional phoneme-independent or phoneme-dependent codebooks

Keywords

hidden Markov models; speech recognition; vector quantisation; HMM parameters; HMM state segments; LVQ2; MKM algorithm; VQ codebook design algorithm; discrimination; hidden Markov model; input speech spectrum; phoneme; quantization distortions; recognition accuracy; recognition units; vector quantization; Algorithm design and analysis; Books; Clustering algorithms; Design methodology; Hidden Markov models; Iterative algorithms; Signal processing; Speech processing; Speech recognition; Stochastic processes;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1992. ICASSP-92., 1992 IEEE International Conference on

Conference_Location

San Francisco, CA

ISSN

1520-6149

Print_ISBN

0-7803-0532-9

Type

conf

DOI

10.1109/ICASSP.1992.225898

Filename

225898