Improvements to PLDA i-vector scoring for short segments clustering

Itay Salmun; Irit Opher; Itshak Lapidot

doi:10.1109/ICSEE.2016.7806108

Improvements to PLDA i-vector scoring for short segments clustering

Itay Salmun
, Irit Opher
, Itshak Lapidot

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

3 Scopus citations

Abstract

This paper extends upon previous work using Mean Shift algorithm to perform speaker clustering on i-vectors generated from short speech segments. In this paper we examine the effectiveness of Spherical Normalization in the presence of different numbers of speakers. This normalization method is not only easy to implement but also improves clustering results. The main improvement is that the proposed mean shift algorithm is more robust to changes in the number of speakers. In the case of 30 speakers, we achieved evaluation parameter K of 75.3 compared to 72.1 with the baseline system.

Original language	English
Title of host publication	2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016
Publisher	Institute of Electrical and Electronics Engineers
ISBN (Electronic)	9781509021529
DOIs	https://doi.org/10.1109/ICSEE.2016.7806108
State	Published - 4 Jan 2017
Externally published	Yes
Event	2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016 - Eilat, Israel Duration: 16 Nov 2016 → 18 Nov 2016

Publication series

Name	2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016

Conference

Conference	2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016
Country/Territory	Israel
City	Eilat
Period	16/11/16 → 18/11/16

Keywords

Probabilistic Linear Discriminant Analysis
Speaker clustering
i-vectors
mean shift clustering
short segments

ASJC Scopus subject areas

Computer Science Applications
Hardware and Architecture
Artificial Intelligence
Computer Networks and Communications
Electrical and Electronic Engineering

Access to Document

10.1109/ICSEE.2016.7806108

Cite this

Salmun, I., Opher, I., & Lapidot, I. (2017). Improvements to PLDA i-vector scoring for short segments clustering. In 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016 Article 7806108 (2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016). Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICSEE.2016.7806108

@inproceedings{8a772628755344b7be3e661ff8ac8fa7,

title = "Improvements to PLDA i-vector scoring for short segments clustering",

abstract = "This paper extends upon previous work using Mean Shift algorithm to perform speaker clustering on i-vectors generated from short speech segments. In this paper we examine the effectiveness of Spherical Normalization in the presence of different numbers of speakers. This normalization method is not only easy to implement but also improves clustering results. The main improvement is that the proposed mean shift algorithm is more robust to changes in the number of speakers. In the case of 30 speakers, we achieved evaluation parameter K of 75.3 compared to 72.1 with the baseline system.",

keywords = "Probabilistic Linear Discriminant Analysis, Speaker clustering, i-vectors, mean shift clustering, short segments",

author = "Itay Salmun and Irit Opher and Itshak Lapidot",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.; 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016 ; Conference date: 16-11-2016 Through 18-11-2016",

year = "2017",

month = jan,

day = "4",

doi = "10.1109/ICSEE.2016.7806108",

language = "English",

series = "2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016",

publisher = "Institute of Electrical and Electronics Engineers",

booktitle = "2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016",

address = "United States",

}

Salmun, I, Opher, I & Lapidot, I 2017, Improvements to PLDA i-vector scoring for short segments clustering. in 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016., 7806108, 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016, Institute of Electrical and Electronics Engineers, 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016, Eilat, Israel, 16/11/16. https://doi.org/10.1109/ICSEE.2016.7806108

Improvements to PLDA i-vector scoring for short segments clustering. / Salmun, Itay; Opher, Irit; Lapidot, Itshak.
2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016. Institute of Electrical and Electronics Engineers, 2017. 7806108 (2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016).

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

TY - GEN

T1 - Improvements to PLDA i-vector scoring for short segments clustering

AU - Salmun, Itay

AU - Opher, Irit

AU - Lapidot, Itshak

PY - 2017/1/4

Y1 - 2017/1/4

N2 - This paper extends upon previous work using Mean Shift algorithm to perform speaker clustering on i-vectors generated from short speech segments. In this paper we examine the effectiveness of Spherical Normalization in the presence of different numbers of speakers. This normalization method is not only easy to implement but also improves clustering results. The main improvement is that the proposed mean shift algorithm is more robust to changes in the number of speakers. In the case of 30 speakers, we achieved evaluation parameter K of 75.3 compared to 72.1 with the baseline system.

AB - This paper extends upon previous work using Mean Shift algorithm to perform speaker clustering on i-vectors generated from short speech segments. In this paper we examine the effectiveness of Spherical Normalization in the presence of different numbers of speakers. This normalization method is not only easy to implement but also improves clustering results. The main improvement is that the proposed mean shift algorithm is more robust to changes in the number of speakers. In the case of 30 speakers, we achieved evaluation parameter K of 75.3 compared to 72.1 with the baseline system.

KW - Probabilistic Linear Discriminant Analysis

KW - Speaker clustering

KW - i-vectors

KW - mean shift clustering

KW - short segments

UR - https://www.scopus.com/pages/publications/85014280407

U2 - 10.1109/ICSEE.2016.7806108

DO - 10.1109/ICSEE.2016.7806108

M3 - Conference contribution

AN - SCOPUS:85014280407

T3 - 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016

BT - 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016

PB - Institute of Electrical and Electronics Engineers

T2 - 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016

Y2 - 16 November 2016 through 18 November 2016

ER -

Salmun I, Opher I, Lapidot I. Improvements to PLDA i-vector scoring for short segments clustering. In 2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016. Institute of Electrical and Electronics Engineers. 2017. 7806108. (2016 IEEE International Conference on the Science of Electrical Engineering, ICSEE 2016). doi: 10.1109/ICSEE.2016.7806108

Improvements to PLDA i-vector scoring for short segments clustering

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this