Publication detail

Towards Lower Error Rates in Phoneme Recognition

SCHWARZ, P., MATĚJKA, P., ČERNOCKÝ, J.

Original Title

Towards Lower Error Rates in Phoneme Recognition

English Title

Towards Lower Error Rates in Phoneme Recognition

Type

conference paper

Language

en

Original Abstract

We investigate techniques for acoustic modeling in automatic recognition of context-independent phoneme strings from the TIMIT database. The baseline phoneme recognizer is based on TempoRAl Patterns (TRAP). This recognizer is simplified to shorten processing times and reduce computational requirements. More states per phoneme and bi-gram language models are incorporated into the system and evaluated. The question of insufficient amount of training data is discussed and the system is improved. All modifications lead to a faster system with about 23.6% relative improvement over the baseline in phoneme error rate.

English abstract

We investigate techniques for acoustic modeling in automatic recognition of context-independent phoneme strings from the TIMIT database. The baseline phoneme recognizer is based on TempoRAl Patterns (TRAP). This recognizer is simplified to shorten processing times and reduce computational requirements. More states per phoneme and bi-gram language models are incorporated into the system and evaluated. The question of insufficient amount of training data is discussed and the system is improved. All modifications lead to a faster system with about 23.6% relative improvement over the baseline in phoneme error rate.

Keywords

phoneme recognition, traps, speech recognition, feature extraction

RIV year

2004

Released

08.09.2004

Publisher

Springer Verlag

Location

Brno

ISBN

3-540-23049-1

Book

Proceedings of 7th International Conference Text,Speech and Dialoque 2004

Pages from

465

Pages to

472

Pages count

8

URL

Documents

BibTex


@inproceedings{BUT17585,
  author="Petr {Schwarz} and Pavel {Matějka} and Jan {Černocký}",
  title="Towards Lower Error Rates in Phoneme Recognition",
  annote="We investigate techniques for acoustic modeling in automatic
recognition of context-independent phoneme strings from the TIMIT
database. The baseline phoneme recognizer is based on TempoRAl Patterns
(TRAP). This recognizer is simplified to shorten processing times and
reduce computational requirements. More states per phoneme and bi-gram
language models are incorporated into the system and evaluated. The
question of insufficient amount of training data is discussed and the
system is improved. All modifications lead to a faster system with
about 23.6% relative improvement over the baseline in phoneme error
rate.",
  address="Springer Verlag",
  booktitle="Proceedings of 7th International Conference Text,Speech and Dialoque 2004",
  chapter="17585",
  institution="Springer Verlag",
  year="2004",
  month="september",
  pages="465",
  publisher="Springer Verlag",
  type="conference paper"
}