Convolutional Neural Network for speaker change detection in telephone speaker diarization system

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F17%3A43932069" target="_blank" >RIV/49777513:23520/17:43932069 - isvavai.cz</a>

Výsledek na webu

<a href="http://ieeexplore.ieee.org/document/7953097/" target="_blank" >http://ieeexplore.ieee.org/document/7953097/</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICASSP.2017.7953097" target="_blank" >10.1109/ICASSP.2017.7953097</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Convolutional Neural Network for speaker change detection in telephone speaker diarization system

Popis výsledku v původním jazyce

The aim of this paper is to propose a speaker change detection technique based on Convolutional Neural Network (CNN) and evaluate its contribution to the performance of a speaker diarization system for telephone conversations. For the comparison we used an i-vector based speaker diarization system. The baseline speaker change detection uses Generalized Likelihood Ratio (GLR) metric. Experiments were conducted on the English part of the CallHome corpus. Our proposed CNN speaker change detection outperformed the GLR approach, reducing the Equal Error Rate relatively by 46 %. The final results on speaker diarization system indicate that the use of speaker change detection based on CNN is beneficial with relative improvement of diarization error rate by 28 %.

Název v anglickém jazyce

Convolutional Neural Network for speaker change detection in telephone speaker diarization system

Popis výsledku anglicky

The aim of this paper is to propose a speaker change detection technique based on Convolutional Neural Network (CNN) and evaluate its contribution to the performance of a speaker diarization system for telephone conversations. For the comparison we used an i-vector based speaker diarization system. The baseline speaker change detection uses Generalized Likelihood Ratio (GLR) metric. Experiments were conducted on the English part of the CallHome corpus. Our proposed CNN speaker change detection outperformed the GLR approach, reducing the Equal Error Rate relatively by 46 %. The final results on speaker diarization system indicate that the use of speaker change detection based on CNN is beneficial with relative improvement of diarization error rate by 28 %.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20205 - Automation and control systems

Projekt

<a href="/cs/project/GBP103%2F12%2FG084" target="_blank" >GBP103/12/G084: Centrum pro multi-modální interpretaci dat velkého rozsahu</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2017

Kód důvěrnosti údajů

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

Název statě ve sborníku

Acoustics, Speech and Signal Processing (ICASSP), 2017 IEEE International Conference on

ISBN

978-1-5090-4117-6

ISSN

1520-6149

e-ISSN

neuvedeno

Počet stran výsledku

5

Strana od-do

4945-4949

Název nakladatele

IEEE Signal Processing Society

Místo vydání

New York

Místo konání akce

New Orleans, LA, USA

Datum konání akce

5. 3. 2017

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000414286205021