CLUSTER CRITERION FUNCTIONS IN SPECTRAL SUBSPACE AND THEIR APPLICATION IN SPEAKER CLUSTERING

被引：4

作者：

Nguyen, Trung Hieu ^{[1
]}

Li, Haizhou ^{[1
]}

Chng, Eng Siong ^{[2
]}

机构：

[1] Inst Infocomm Res, Dept Human Language Technol, 1 Fusionopolis Way,21-01 Connexis,South Tower, Singapore 138632, Singapore

[2] Nanyang Technol Univ, Sch Comp Engn, Singapore 639798, Singapore

来源：

2009 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1- 8, PROCEEDINGS | 2009年

关键词：

speaker diarization; criterion function; spectral clustering;

D O I：

10.1109/ICASSP.2009.4960526

中图分类号：

O42 [声学];

学科分类号：

070206 ; 082403 ;

摘要：

In this paper, we propose two cluster criterion functions which aim to maximize the separation between intra-cluster distances and inter-cluster distances. These criteria can automatically deduce the desired number of clusters based on their extremized values. We then propose an algorithm to apply our criterion functions in conjunction with spectral clustering. By exploiting the characteristic of spectral subspace,we show that the speakers are more separable in this subspace which will further enhance the effectiveness of our proposed criteria. The algorithm is used in our agglomerative hierarchical speaker diarization system to test on Rich Transcription 2007 conference data set and obtains very good results.

引用

页码：4085 / +

页数：2

共 10 条

[1] A robust speaker clustering algorithm [J].

Ajmera, J ;

Wooters, C .

ASRU'03: 2003 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING ASRU '03, 2003, :411-416

[2]

[Anonymous], 1997, 5 EUR C SPEECH COMM

[3]

[Anonymous], 2002, ADV NEURAL INFORM PR

[4]

Conover William Jay, 1999, Practical nonparametric statistics, V350

[5]

Fanti C, 2004, ADV NEUR IN, V16, P1603

[6]

Jin H., 1997, PROC, P108

[7]

*NIST, SPRING 2007 RT 07 RI

[8]

*NIST, RICH TRANSCR 2007 SP

[9]

Siegler M.A., 1997, P DARPA SPEECH REC W, V1997

[10]

VANLEEUWEN D, 2005, NIST 2005 SPRING RIC

← 1 →