Analysis of DNN Speech Signal Enhancement for Robust Speaker Recognition

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F19%3APU132989" target="_blank" >RIV/00216305:26230/19:PU132989 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.sciencedirect.com/science/article/pii/S0885230818303607" target="_blank" >https://www.sciencedirect.com/science/article/pii/S0885230818303607</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.csl.2019.06.004" target="_blank" >10.1016/j.csl.2019.06.004</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Analysis of DNN Speech Signal Enhancement for Robust Speaker Recognition

Popis výsledku v původním jazyce

In this work, we present an analysis of a DNN-based autoencoder for speech enhancement, dereverberation and denoising. Thetarget application is a robust speaker verification (SV) system. We start our approach by carefully designing a data augmentationprocess to cover a wide range of acoustic conditions and to obtain rich training data for various components of our SV system.We augment several well-known databases used in SV with artificially noised and reverberated data and we use them to train adenoising autoencoder (mapping noisy and reverberated speech to its clean version) as well as an x-vector extractor which is cur-rently considered as state-of-the-art in SV. Later, we use the autoencoder as a preprocessing step for a text-independent SV sys-tem. We compare results achieved with autoencoder enhancement, multi-condition PLDA training and their simultaneous use.We present a detailed analysis with various conditions of NIST SRE 2010, 2016, PRISM and with re-transmitted data. We con-clude that the proposed preprocessing can significantly improve both i-vector and x-vector baselines and that this technique canbe used to build a robust SV system for various target domains.

Název v anglickém jazyce

Analysis of DNN Speech Signal Enhancement for Robust Speaker Recognition

Popis výsledku anglicky

In this work, we present an analysis of a DNN-based autoencoder for speech enhancement, dereverberation and denoising. Thetarget application is a robust speaker verification (SV) system. We start our approach by carefully designing a data augmentationprocess to cover a wide range of acoustic conditions and to obtain rich training data for various components of our SV system.We augment several well-known databases used in SV with artificially noised and reverberated data and we use them to train adenoising autoencoder (mapping noisy and reverberated speech to its clean version) as well as an x-vector extractor which is cur-rently considered as state-of-the-art in SV. Later, we use the autoencoder as a preprocessing step for a text-independent SV sys-tem. We compare results achieved with autoencoder enhancement, multi-condition PLDA training and their simultaneous use.We present a detailed analysis with various conditions of NIST SRE 2010, 2016, PRISM and with re-transmitted data. We con-clude that the proposed preprocessing can significantly improve both i-vector and x-vector baselines and that this technique canbe used to build a robust SV system for various target domains.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

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)

Rok uplatnění

2019

Kód důvěrnosti údajů

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

Název periodika

COMPUTER SPEECH AND LANGUAGE

ISSN

0885-2308

e-ISSN

1095-8363

Svazek periodika

2019

Číslo periodika v rámci svazku

58

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

19

Strana od-do

403-421

Kód UT WoS článku

000477663800022

EID výsledku v databázi Scopus

2-s2.0-85067550556