Blind Extraction of Target Speech Source: Three ways of Guidance Exploiting Supervised Speaker Embeddings

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F46747885%3A24220%2F22%3A00009854" target="_blank" >RIV/46747885:24220/22:00009854 - isvavai.cz</a>

Výsledek na webu

<a href="https://asap.ite.tul.cz/wp-content/uploads/sites/3/2022/08/paper_16.pdf" target="_blank" >https://asap.ite.tul.cz/wp-content/uploads/sites/3/2022/08/paper_16.pdf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/IWAENC53105.2022.9914778" target="_blank" >10.1109/IWAENC53105.2022.9914778</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Blind Extraction of Target Speech Source: Three ways of Guidance Exploiting Supervised Speaker Embeddings

Popis výsledku v původním jazyce

The manuscript deals with the robust extraction of a speaker of interest (SOI) from a mixture of audio sources. A blind algorithm based on independent vector extraction (IVE) is used, which, by definition, extracts an arbitrary source. To focus the extraction towards the SOI, a prior knowledge identifying the target source is required. To this end, the manuscript exploits speaker-identification based on embedding features computed via a pretrained forward sequential memory network (FSMN). We introduce and experimentally validate three ways how this prior knowledge can be employed in a blind algorithm, namely, speaker-specific initialization, pilot signal, and supervised deflation of the mixture. The experiments show that the proposed techniques complement each other and lead to robust identification/extraction of the SOI in difficult mixtures of three speakers.

Název v anglickém jazyce

Blind Extraction of Target Speech Source: Three ways of Guidance Exploiting Supervised Speaker Embeddings

Popis výsledku anglicky

The manuscript deals with the robust extraction of a speaker of interest (SOI) from a mixture of audio sources. A blind algorithm based on independent vector extraction (IVE) is used, which, by definition, extracts an arbitrary source. To focus the extraction towards the SOI, a prior knowledge identifying the target source is required. To this end, the manuscript exploits speaker-identification based on embedding features computed via a pretrained forward sequential memory network (FSMN). We introduce and experimentally validate three ways how this prior knowledge can be employed in a blind algorithm, namely, speaker-specific initialization, pilot signal, and supervised deflation of the mixture. The experiments show that the proposed techniques complement each other and lead to robust identification/extraction of the SOI in difficult mixtures of three speakers.

Druh

D - Stať ve sborníku

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

2022

Kód důvěrnosti údajů

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

Název statě ve sborníku

International Workshop on Acoustic Signal Enhancement

ISBN

978-166546867-1

ISSN

—

e-ISSN

—

Počet stran výsledku

5

Strana od-do

—

Název nakladatele

IEEE

Místo vydání

—

Místo konání akce

Bamberg

Datum konání akce

1. 1. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000934046400073