Source Separation for Sound Event Detection in domestic environments using jointly trained models

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F22%3APU146145" target="_blank" >RIV/00216305:26230/22:PU146145 - isvavai.cz</a>

Výsledek na webu

<a href="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9914755" target="_blank" >https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9914755</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/IWAENC53105.2022.9914755" target="_blank" >10.1109/IWAENC53105.2022.9914755</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Source Separation for Sound Event Detection in domestic environments using jointly trained models

Popis výsledku v původním jazyce

Sound Event Detection and Source Separation are closely related tasks: whereas the first aims to find the time boundaries of acoustic events inside a recording, the goal of the latter is to isolate each of the acoustic sources into different signals. This paper presents a Sound Event Detection system formed by two independently pretrained blocks for Source Separation and Sound Event Detection. We propose a joint-training scheme, where both blocks are trained at the same time, and a two-stage training, where each block trains while the other one is frozen. In addition, we compare the use of supervised and unsupervised pre-training for the Separation block, and two model selection strategies for Sound Event Detection. Our experiments show that the proposed methods are able to outperform the baseline systems of the DCASE 2021 Challenge Task 4.

Název v anglickém jazyce

Source Separation for Sound Event Detection in domestic environments using jointly trained models

Popis výsledku anglicky

Sound Event Detection and Source Separation are closely related tasks: whereas the first aims to find the time boundaries of acoustic events inside a recording, the goal of the latter is to isolate each of the acoustic sources into different signals. This paper presents a Sound Event Detection system formed by two independently pretrained blocks for Source Separation and Sound Event Detection. We propose a joint-training scheme, where both blocks are trained at the same time, and a two-stage training, where each block trains while the other one is frozen. In addition, we compare the use of supervised and unsupervised pre-training for the Separation block, and two model selection strategies for Sound Event Detection. Our experiments show that the proposed methods are able to outperform the baseline systems of the DCASE 2021 Challenge Task 4.

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

<a href="/cs/project/LTAIN19087" target="_blank" >LTAIN19087: Multi-lingualita v řečových technologiích</a><br>

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

Proceedings of The 17th International Workshop on Acoustic Signal Enhancement (IWAENC 2022)

ISBN

978-1-6654-6867-1

ISSN

—

e-ISSN

—

Počet stran výsledku

5

Strana od-do

1-5

Název nakladatele

IEEE Signal Processing Society

Místo vydání

Bamberg

Místo konání akce

Bamberg

Datum konání akce

5. 9. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—