A Spearman correlation coefficient ranking for matching-score fusion on speaker recognition

Di Liu, Siu Yeung Cho, Dong Mei Sun, Zheng Ding Qiu

Research output: Chapter in Book/Conference proceedingConference contributionpeer-review

26 Citations (Scopus)

Abstract

This paper describes an attempt by making use of the Spearman rank correlation coefficient as a metric to measure features on the matching score fusion of speaker recognition. In the context of fusion technique of speaker recognition, the Spearman coefficient is introduced to measure the correlation between different acoustic features combining together such that the metric is able to find out an optimal solution for selecting a set of feasible features to achieve good performance. This coefficient can evaluate how far the relationship between the combined features in term of scores combined by two features. Throughout the evaluations for the scores combined by obtaining by different combination of acoustic features, we found that MFCC and residual phase are the optimal solution for feature selection. The outcomes indicate that the Spearman correlation coefficient is a reliable metric to measure features in the fusion of speaker recognition.

Original languageEnglish
Title of host publicationTENCON 2010 - 2010 IEEE Region 10 Conference
Pages736-741
Number of pages6
DOIs
Publication statusPublished - 2010
Externally publishedYes
Event2010 IEEE Region 10 Conference, TENCON 2010 - Fukuoka, Japan
Duration: 21 Nov 201024 Nov 2010

Publication series

NameIEEE Region 10 Annual International Conference, Proceedings/TENCON

Conference

Conference2010 IEEE Region 10 Conference, TENCON 2010
Country/TerritoryJapan
CityFukuoka
Period21/11/1024/11/10

Keywords

  • Matching-score level fusion
  • Speaker recognition
  • Spearman rank correlation coefficient

ASJC Scopus subject areas

  • Computer Science Applications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'A Spearman correlation coefficient ranking for matching-score fusion on speaker recognition'. Together they form a unique fingerprint.

Cite this