Telephone Speech Recognition Using Time-Domain IIR Filter Bank in MFCC Computation

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F15%3A00230533" target="_blank" >RIV/68407700:21230/15:00230533 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Telephone Speech Recognition Using Time-Domain IIR Filter Bank in MFCC Computation

Popis výsledku v původním jazyce

This paper analyses the influence of two different methods of MFCC features computation according to resulting accuracy within the telephone speech recognition. These methods include the standard frequency based approach and computation via the time-domain IIR filter bank. Also, the influences of the amplitude and surface normalization of the filterbank were examined. These parametrizations were tested on multilingual telephone Speechdat database, so our effort was to unify the recognition process across 5 languages - Russian, Slovak, Czech, Polish and Hungarian. Best obtained results for individual languages were PER of 27.29 % for Czech, 30.64 % for Slovak, 25.52 % for Hungarian, 34.81 % for Polish and 39.58 % for Russian.

Název v anglickém jazyce

Telephone Speech Recognition Using Time-Domain IIR Filter Bank in MFCC Computation

Popis výsledku anglicky

This paper analyses the influence of two different methods of MFCC features computation according to resulting accuracy within the telephone speech recognition. These methods include the standard frequency based approach and computation via the time-domain IIR filter bank. Also, the influences of the amplitude and surface normalization of the filterbank were examined. These parametrizations were tested on multilingual telephone Speechdat database, so our effort was to unify the recognition process across 5 languages - Russian, Slovak, Czech, Polish and Hungarian. Best obtained results for individual languages were PER of 27.29 % for Czech, 30.64 % for Slovak, 25.52 % for Hungarian, 34.81 % for Polish and 39.58 % for Russian.

Druh

D - Stať ve sborníku

CEP obor

JA - Elektronika a optoelektronika, elektrotechnika

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2015

Kód důvěrnosti údajů

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

Název statě ve sborníku

Proceedings of the 19th International Scientific Student Conferenece POSTER 2015

ISBN

978-80-01-05499-4

ISSN

—

e-ISSN

—

Počet stran výsledku

4

Strana od-do

1-4

Název nakladatele

Czech Technical University in Prague

Místo vydání

Praha

Místo konání akce

Praha

Datum konání akce

14. 5. 2015

Typ akce podle státní příslušnosti

EUR - Evropská akce

Kód UT WoS článku

—