Improving of LVCSR for Casual Czech Using Publicly Available Language Resources

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F17%3A00313500" target="_blank" >RIV/68407700:21230/17:00313500 - isvavai.cz</a>

Výsledek na webu

<a href="https://link.springer.com/chapter/10.1007/978-3-319-66429-3_42" target="_blank" >https://link.springer.com/chapter/10.1007/978-3-319-66429-3_42</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-319-66429-3_42" target="_blank" >10.1007/978-3-319-66429-3_42</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Improving of LVCSR for Casual Czech Using Publicly Available Language Resources

Popis výsledku v původním jazyce

The paper presents the design of Czech casual speech recognition which is a part of the wider research focused on understanding very informal speaking styles. The study was carried out using the NCCCz corpus and the contributions of optimized acoustic and language models as well as pronunciation lexicon optimization were analyzed. Special attention was paid to the impact of publicly available corpora suitable for language model (LM) creation. Our final DNN-HMM system achieved in the task of casual speech recognition WER of 30-60% depending on LM used. The results of recognition for other speaking styles are presented as well for the comparison purposes. The system was built using KALDI toolkit and created recipes are available for the research community.

Název v anglickém jazyce

Improving of LVCSR for Casual Czech Using Publicly Available Language Resources

Popis výsledku anglicky

The paper presents the design of Czech casual speech recognition which is a part of the wider research focused on understanding very informal speaking styles. The study was carried out using the NCCCz corpus and the contributions of optimized acoustic and language models as well as pronunciation lexicon optimization were analyzed. Special attention was paid to the impact of publicly available corpora suitable for language model (LM) creation. Our final DNN-HMM system achieved in the task of casual speech recognition WER of 30-60% depending on LM used. The results of recognition for other speaking styles are presented as well for the comparison purposes. The system was built using KALDI toolkit and created recipes are available for the research community.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20202 - Communication engineering and systems

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2017

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Speech and Computer

ISBN

978-3-319-66428-6

ISSN

0302-9743

e-ISSN

—

Počet stran výsledku

11

Strana od-do

427-437

Název nakladatele

Springer

Místo vydání

Heidelberg

Místo konání akce

Hatfield

Datum konání akce

12. 9. 2017

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—