Text-dependent Speaker Recognition in Low-cost Systems

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26220%2F11%3APU94560" target="_blank" >RIV/00216305:26220/11:PU94560 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Text-dependent Speaker Recognition in Low-cost Systems

Popis výsledku v původním jazyce

Although the text-dependent recognition is not as universal as the text-independent recognition it has still many applications. Moreover the accuracy of the systems based on the text-dependent recognition is still higher than in case of the text-independent systems. This work proposes new techniques in the field of text-dependent speaker recognition so that it is possible to implement these systems in low-cost devices. It introduces a feature vector called voice imprint, which has low memory requirements, and a classifier based on fractional distances which is very fast and accurate in comparison with another classifiers. It is shown that using the fractional distances along with the voice imprint, it is possible to reach successful detection rate equal to 96.94 % and 82.29 % in case of microphone mismatch. During verification, system reached equal error rate 3.93 % and 10.43 % when assuming microphone mismatch. System was tested using database which consists of 48 speakers (three 3 s

Název v anglickém jazyce

Text-dependent Speaker Recognition in Low-cost Systems

Popis výsledku anglicky

Although the text-dependent recognition is not as universal as the text-independent recognition it has still many applications. Moreover the accuracy of the systems based on the text-dependent recognition is still higher than in case of the text-independent systems. This work proposes new techniques in the field of text-dependent speaker recognition so that it is possible to implement these systems in low-cost devices. It introduces a feature vector called voice imprint, which has low memory requirements, and a classifier based on fractional distances which is very fast and accurate in comparison with another classifiers. It is shown that using the fractional distances along with the voice imprint, it is possible to reach successful detection rate equal to 96.94 % and 82.29 % in case of microphone mismatch. During verification, system reached equal error rate 3.93 % and 10.43 % when assuming microphone mismatch. System was tested using database which consists of 48 speakers (three 3 s

Druh

D - Stať ve sborníku

CEP obor

JA - Elektronika a optoelektronika, elektrotechnika

OECD FORD obor

—

Projekt

Výsledek vznikl pri realizaci vícero projektů. Více informací v záložce Projekty.

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>Z - Vyzkumny zamer (s odkazem do CEZ)<br>S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2011

Kód důvěrnosti údajů

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

Název statě ve sborníku

6th International Conference on Teleinformatics

ISBN

978-80-214-4231-3

ISSN

—

e-ISSN

—

Počet stran výsledku

5

Strana od-do

154-158

Název nakladatele

Neuveden

Místo vydání

Neuveden

Místo konání akce

Dolní Morava

Datum konání akce

2. 2. 2011

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—