Publication detail

Data Driven Design of Filter Bank for Speech Recognition

BURGET, L. HERMANSKY, H.

English title

Data Driven Design of Filter Bank for Speech Recognition

Type

conference paper

Language

en

Original abstract

Filter bank approach is commonly used in feature extraction phase of speech recognition (e.g. Mel frequency cepstral coefficients). Filter bank is applied for modification of magnitude spectrum according to physiological and psychological findings. However, since mechanism of human auditory system is not fully understood,the optimal filter bank parameters are not known. This work presents a method where the filter bank, optimized for discriminability between phonemes, is derived directly from phonetically labeled speech data using Linear Discriminant Analysis. This work can be seen as another proof of the fact that incorporation of psychoacoustic findings into feature extraction can lead to better recognition performance.

English abstract

Filter bank approach is commonly used in feature extraction phase of speech recognition (e.g. Mel frequency cepstral coefficients). Filter bank is applied for modification of magnitude spectrum according to physiological and psychological findings. However, since mechanism of human auditory system is not fully understood,the optimal filter bank parameters are not known. This work presents a method where the filter bank, optimized for discriminability between phonemes, is derived directly from phonetically labeled speech data using Linear Discriminant Analysis. This work can be seen as another proof of the fact that incorporation of psychoacoustic findings into feature extraction can lead to better recognition performance.

Keywords in English

mel filterbank, speech recognition, linear discriminant analysis

RIV year

2001

Released

01.09.2001

Publisher

Springer Verlag

Location

Zelezna Ruda

ISBN

3-540-42557-8

Book

Proc. 4th Intl. Conference Text, Speech Dialogue

Pages count

6

BIBTEX


@inproceedings{BUT3682,
  author="Lukáš {Burget} and Hynek {Hermansky},
  title="Data Driven Design of Filter Bank for Speech Recognition",
  booktitle="Proc. 4th Intl. Conference Text, Speech Dialogue",
  year="2001",
  month="September",
  publisher="Springer Verlag",
  address="Zelezna Ruda",
  isbn="3-540-42557-8"
}