Scalar quantization of features in discrete hidden Markov models

Author

Ramachandrula, Sitaram ; Thippur, Sreenivas

Author_Institution

Dept. of Electr. Commun. Eng., Indian Inst. of Sci., Bangalore, India

Volume

1

fYear

1997

fDate

9-12 Sep 1997

Firstpage

537

Abstract

Traditionally, discrete hidden Markov models (D-HMM) use vector quantized speech feature vectors. In this paper, we propose scalar quantization of each element of the speech feature vector in the D-HMM formulation. The alteration required in the D-HMM algorithms for this modification is discussed here. A comparison is made between the performance of D-HMM based speech recognizers using scalar and vector quantization of speech features respectively. A speaker independent TIMIT vowel classification experiment is chosen for this task. It is observed that the scalar quantization of features enhances the vowel classification accuracy by 8 to 9%, compared to VQ based D-HMM. Also, the number of HMM parameters to estimate from a given amount of training data is drastically reduced

Keywords

hidden Markov models; pattern classification; quantisation (signal); speech recognition; D-HMM algorithms; D-HMM based speech recognizers; discrete hidden Markov models; scalar quantization; speaker independent TIMIT vowel classification experiment; speech feature vectors; speech features; speech recognition; vector quantization; Cepstral analysis; Hidden Markov models; Parameter estimation; Probability density function; Real time systems; Spatial databases; Speech recognition; Table lookup; Training data; Vector quantization;

fLanguage

English

Publisher

ieee

Conference_Titel

Information, Communications and Signal Processing, 1997. ICICS., Proceedings of 1997 International Conference on

Print_ISBN

0-7803-3676-3

Type

conf

DOI

10.1109/ICICS.1997.647156

Filename

647156