Detail publikace

SpeakerBeam: A New Deep Learning Technology for Extracting Speech of a Target Speaker Based on the Speaker's Voice Characteristics

DELCROIX, M. ŽMOLÍKOVÁ, K. KINOSHITA, K. ARAKI, S. OGAWA, A. NAKATANI, T.

Originální název

SpeakerBeam: A New Deep Learning Technology for Extracting Speech of a Target Speaker Based on the Speaker's Voice Characteristics

Typ

článek v časopise ve Scopus, Jsc

Jazyk

angličtina

Originální abstrakt

In a noisy environment such as a cocktail party, humans can focus on listening to a desired speaker, an ability known as selective hearing. Current approaches developed to realize computational selective hearing require knowing the position of the target speaker, which limits their practical usage. This article introduces SpeakerBeam, a deep learning based approach for computational selective hearing based on the characteristics of the target speakers voice. SpeakerBeam requires only a small amount of speech data from the target speaker to compute his/her voice characteristics. It can then extract the speech of that speaker regardless of his/her position or the number of speakers talking in the background.

Klíčová slova

deep learning, target speaker extraction, SpeakerBeam

Autoři

DELCROIX, M.; ŽMOLÍKOVÁ, K.; KINOSHITA, K.; ARAKI, S.; OGAWA, A.; NAKATANI, T.

Vydáno

1. 11. 2018

ISSN

1348-3447

Periodikum

NTT Technical Review

Ročník

16

Číslo

11

Stát

Japonsko

Strany od

19

Strany do

24

Strany počet

6

URL

https://www.ntt-review.jp/archive/ntttechnical.php?contents=ntr201811all.pdf&mode=show_pdf

BibTex

@article{BUT185149,
  author="DELCROIX, M. and ŽMOLÍKOVÁ, K. and KINOSHITA, K. and ARAKI, S. and OGAWA, A. and NAKATANI, T.",
  title="SpeakerBeam: A New Deep Learning Technology for Extracting Speech of a Target Speaker Based on the Speaker's Voice Characteristics",
  journal="NTT Technical Review",
  year="2018",
  volume="16",
  number="11",
  pages="19--24",
  issn="1348-3447",
  url="https://www.ntt-review.jp/archive/ntttechnical.php?contents=ntr201811all.pdf&mode=show_pdf"
}