Robust Recognition of Speech with Background Music in Acoustically Under-Resourced Scenarios

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F46747885%3A24220%2F18%3A00006123" target="_blank" >RIV/46747885:24220/18:00006123 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1109/ICASSP.2018.8462674" target="_blank" >http://dx.doi.org/10.1109/ICASSP.2018.8462674</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICASSP.2018.8462674" target="_blank" >10.1109/ICASSP.2018.8462674</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Robust Recognition of Speech with Background Music in Acoustically Under-Resourced Scenarios

Popis výsledku v původním jazyce

This paper addresses the task of Automatic Speech Recognition (ASR) with music in the background. We consider two different situations: 1) scenarios with very small amount of labeled training utterances (duration 1 hour) and 2) scenarios with large amount of labeled training utterances (duration 132 hours). In these situations, we aim to achieve robust recognition. To this end we investigate the following techniques: a) multi-condition training of the acoustic model, b) denoising autoencoders for feature enhancement and c) joint training of both above mentioned techniques. We demonstrate that the considered methods can be successfully trained with the small amount of labeled acoustic data. We present substantially improved performance compared to acoustic models trained on clean speech. Further, we show a significant increase of accuracy in the under-resourced scenario, when utilizing additional amount of non-labeled data. Here, the non-labeled dataset is used to improve the accuracy of the feature enhancement via autoencoders. Subsequently, the autoencoders are jointly fine-tuned along with the acoustic model using the small amount of labeled utterances.

Název v anglickém jazyce

Robust Recognition of Speech with Background Music in Acoustically Under-Resourced Scenarios

Popis výsledku anglicky

This paper addresses the task of Automatic Speech Recognition (ASR) with music in the background. We consider two different situations: 1) scenarios with very small amount of labeled training utterances (duration 1 hour) and 2) scenarios with large amount of labeled training utterances (duration 132 hours). In these situations, we aim to achieve robust recognition. To this end we investigate the following techniques: a) multi-condition training of the acoustic model, b) denoising autoencoders for feature enhancement and c) joint training of both above mentioned techniques. We demonstrate that the considered methods can be successfully trained with the small amount of labeled acoustic data. We present substantially improved performance compared to acoustic models trained on clean speech. Further, we show a significant increase of accuracy in the under-resourced scenario, when utilizing additional amount of non-labeled data. Here, the non-labeled dataset is used to improve the accuracy of the feature enhancement via autoencoders. Subsequently, the autoencoders are jointly fine-tuned along with the acoustic model using the small amount of labeled utterances.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20206 - Computer hardware and architecture

Projekt

<a href="/cs/project/TH03010018" target="_blank" >TH03010018: DeepSpot - Multilingvální technologie pro detekci a včasné upozornění</a><br>

Návaznosti

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

Rok uplatnění

2018

Kód důvěrnosti údajů

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

Název statě ve sborníku

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

ISBN

978-153864658-8

ISSN

1520-6149

e-ISSN

—

Počet stran výsledku

5

Strana od-do

5624-5628

Název nakladatele

IEEE

Místo vydání

Kanada

Místo konání akce

Calgary, Kanada

Datum konání akce

1. 1. 2018

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000446384605157