A thousand words are worth more than one recording: Word-embedding based speaker change detection

Or Haim Anidjar; Itshak Lapidot; Chen Hajaj; Amit Dvir

doi:10.21437/interspeech.2021-87

A thousand words are worth more than one recording: Word-embedding based speaker change detection

Or Haim Anidjar, Itshak Lapidot, Chen Hajaj, Amit Dvir

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

3 Scopus citations

Abstract

Speaker Change Detection (SCD) is the task of segmenting an input audio-recording according to speaker interchanges. This task is essential for many applications, such as automatic voice transcription or Speaker Diarization (SD). This paper focuses on the essential task of audio segmentation and suggests a word-embedding-based solution for the SCD problem. Moreover, we show how to use our approach in order to outperform voice-based solutions for the SD problem. We empirically show that our method can accurately identify the speaker-turns in an audio-recording with 82.12% and 89.02% success in the Recall and F1-score measures.

Original language	English
Title of host publication	22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021
Publisher	International Speech Communication Association
Pages	2473-2477
Number of pages	5
ISBN (Electronic)	9781713836902
DOIs	https://doi.org/10.21437/interspeech.2021-87
State	Published - 2021
Externally published	Yes
Event	22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021 - Brno, Czech Republic Duration: 30 Aug 2021 → 3 Sep 2021

Publication series

Name	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Volume	4
ISSN (Print)	2308-457X
ISSN (Electronic)	1990-9772

Conference

Conference	22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021
Country/Territory	Czech Republic
City	Brno
Period	30/08/21 → 3/09/21

Bibliographical note

Publisher Copyright:
Copyright © 2021 ISCA.

Funding

The authors wish to thank IFAT Group that provided the dataset for this research. In addition, this work was supported by the Ariel Cyber Innovation Center in conjunction with the Israel National Cyber directorate in the Prime Minister’s Office.

Keywords

Clustering
Speaker change detection
Speaker diarization
Speech recognition
Word embedding

Access to Document

10.21437/interspeech.2021-87

Cite this

Anidjar, O. H., Lapidot, I., Hajaj, C., & Dvir, A. (2021). A thousand words are worth more than one recording: Word-embedding based speaker change detection. In 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021 (pp. 2473-2477). (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH; Vol. 4). International Speech Communication Association. https://doi.org/10.21437/interspeech.2021-87

Anidjar, Or Haim ; Lapidot, Itshak ; Hajaj, Chen et al. / A thousand words are worth more than one recording : Word-embedding based speaker change detection. 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. International Speech Communication Association, 2021. pp. 2473-2477 (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH).

@inproceedings{cea80f5a7b7f434780cfc5873a2ce805,

title = "A thousand words are worth more than one recording: Word-embedding based speaker change detection",

abstract = "Speaker Change Detection (SCD) is the task of segmenting an input audio-recording according to speaker interchanges. This task is essential for many applications, such as automatic voice transcription or Speaker Diarization (SD). This paper focuses on the essential task of audio segmentation and suggests a word-embedding-based solution for the SCD problem. Moreover, we show how to use our approach in order to outperform voice-based solutions for the SD problem. We empirically show that our method can accurately identify the speaker-turns in an audio-recording with 82.12% and 89.02% success in the Recall and F1-score measures.",

keywords = "Clustering, Speaker change detection, Speaker diarization, Speech recognition, Word embedding",

author = "Anidjar, {Or Haim} and Itshak Lapidot and Chen Hajaj and Amit Dvir",

note = "Publisher Copyright: Copyright {\textcopyright} 2021 ISCA.; 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021 ; Conference date: 30-08-2021 Through 03-09-2021",

year = "2021",

doi = "10.21437/interspeech.2021-87",

language = "אנגלית",

series = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

publisher = "International Speech Communication Association",

pages = "2473--2477",

booktitle = "22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021",

}

Anidjar, OH, Lapidot, I, Hajaj, C & Dvir, A 2021, A thousand words are worth more than one recording: Word-embedding based speaker change detection. in 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, vol. 4, International Speech Communication Association, pp. 2473-2477, 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021, Brno, Czech Republic, 30/08/21. https://doi.org/10.21437/interspeech.2021-87

A thousand words are worth more than one recording: Word-embedding based speaker change detection. / Anidjar, Or Haim; Lapidot, Itshak; Hajaj, Chen et al.
22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. International Speech Communication Association, 2021. p. 2473-2477 (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH; Vol. 4).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A thousand words are worth more than one recording

T2 - 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021

AU - Anidjar, Or Haim

AU - Lapidot, Itshak

AU - Hajaj, Chen

AU - Dvir, Amit

PY - 2021

Y1 - 2021

N2 - Speaker Change Detection (SCD) is the task of segmenting an input audio-recording according to speaker interchanges. This task is essential for many applications, such as automatic voice transcription or Speaker Diarization (SD). This paper focuses on the essential task of audio segmentation and suggests a word-embedding-based solution for the SCD problem. Moreover, we show how to use our approach in order to outperform voice-based solutions for the SD problem. We empirically show that our method can accurately identify the speaker-turns in an audio-recording with 82.12% and 89.02% success in the Recall and F1-score measures.

AB - Speaker Change Detection (SCD) is the task of segmenting an input audio-recording according to speaker interchanges. This task is essential for many applications, such as automatic voice transcription or Speaker Diarization (SD). This paper focuses on the essential task of audio segmentation and suggests a word-embedding-based solution for the SCD problem. Moreover, we show how to use our approach in order to outperform voice-based solutions for the SD problem. We empirically show that our method can accurately identify the speaker-turns in an audio-recording with 82.12% and 89.02% success in the Recall and F1-score measures.

KW - Clustering

KW - Speaker change detection

KW - Speaker diarization

KW - Speech recognition

KW - Word embedding

UR - http://www.scopus.com/inward/record.url?scp=85111078902&partnerID=8YFLogxK

U2 - 10.21437/interspeech.2021-87

DO - 10.21437/interspeech.2021-87

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85111078902

T3 - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

SP - 2473

EP - 2477

BT - 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021

PB - International Speech Communication Association

Y2 - 30 August 2021 through 3 September 2021

ER -

Anidjar OH, Lapidot I, Hajaj C, Dvir A. A thousand words are worth more than one recording: Word-embedding based speaker change detection. In 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. International Speech Communication Association. 2021. p. 2473-2477. (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH). doi: 10.21437/interspeech.2021-87

A thousand words are worth more than one recording: Word-embedding based speaker change detection

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this