An adaptive multibiometric system for uncertain audio condition

Dzati Athiar Ramli, Salina Abdul Samad, Aini Hussain

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Performances of speaker verification systems are superb in clean noise-free conditions but the reliability of the systems drop severely in noisy environments. In this study, we propose a novel approach by introducing Support Vector Machine (SVM) as indicator system for audio reliability estimation. This approach directly validate the quality of the incoming (claimant) speech signal so as to adaptively change the weighting factor for fusion of both subsystem scores. The effectiveness of this approach has been experimented to a multibiometric verification system that employs lipreading images as visual features. This verification system uses SVM as a classifier for both subsystems. Principle Component Analysis (PCA) technique is executed for visual features extraction while for the audio feature extraction; Linear Predictive Coding (LPC) technique has been utilized. In this study, we found that the SVM indicator system is able to determine the quality of the speech signal up to 99.66%. At 10 dB SNR, EER performances are observed as 51.13%, 9.3%, and 0.27% for audio only system, fixed weighting system and adaptive weighting system, respectively.

Original languageEnglish
Title of host publicationLecture Notes in Electrical Engineering
Pages165-177
Number of pages13
Volume60 LNEE
DOIs
Publication statusPublished - 2010
EventInternational Conference in Electronic Engineering and Computing Technology - London
Duration: 1 Jul 20093 Jul 2009

Publication series

NameLecture Notes in Electrical Engineering
Volume60 LNEE
ISSN (Print)18761100
ISSN (Electronic)18761119

Other

OtherInternational Conference in Electronic Engineering and Computing Technology
CityLondon
Period1/7/093/7/09

Fingerprint

Adaptive systems
Support vector machines
Feature extraction
Classifiers
Fusion reactions

Keywords

  • Biometric verification system
  • reliability estimation
  • support vector machine

ASJC Scopus subject areas

  • Industrial and Manufacturing Engineering

Cite this

Ramli, D. A., Abdul Samad, S., & Hussain, A. (2010). An adaptive multibiometric system for uncertain audio condition. In Lecture Notes in Electrical Engineering (Vol. 60 LNEE, pp. 165-177). (Lecture Notes in Electrical Engineering; Vol. 60 LNEE). https://doi.org/10.1007/978-90-481-8776-8_15

An adaptive multibiometric system for uncertain audio condition. / Ramli, Dzati Athiar; Abdul Samad, Salina; Hussain, Aini.

Lecture Notes in Electrical Engineering. Vol. 60 LNEE 2010. p. 165-177 (Lecture Notes in Electrical Engineering; Vol. 60 LNEE).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ramli, DA, Abdul Samad, S & Hussain, A 2010, An adaptive multibiometric system for uncertain audio condition. in Lecture Notes in Electrical Engineering. vol. 60 LNEE, Lecture Notes in Electrical Engineering, vol. 60 LNEE, pp. 165-177, International Conference in Electronic Engineering and Computing Technology, London, 1/7/09. https://doi.org/10.1007/978-90-481-8776-8_15
Ramli DA, Abdul Samad S, Hussain A. An adaptive multibiometric system for uncertain audio condition. In Lecture Notes in Electrical Engineering. Vol. 60 LNEE. 2010. p. 165-177. (Lecture Notes in Electrical Engineering). https://doi.org/10.1007/978-90-481-8776-8_15
Ramli, Dzati Athiar ; Abdul Samad, Salina ; Hussain, Aini. / An adaptive multibiometric system for uncertain audio condition. Lecture Notes in Electrical Engineering. Vol. 60 LNEE 2010. pp. 165-177 (Lecture Notes in Electrical Engineering).
@inproceedings{6294fbfafc8c4e7b9fb24311f4cb56ff,
title = "An adaptive multibiometric system for uncertain audio condition",
abstract = "Performances of speaker verification systems are superb in clean noise-free conditions but the reliability of the systems drop severely in noisy environments. In this study, we propose a novel approach by introducing Support Vector Machine (SVM) as indicator system for audio reliability estimation. This approach directly validate the quality of the incoming (claimant) speech signal so as to adaptively change the weighting factor for fusion of both subsystem scores. The effectiveness of this approach has been experimented to a multibiometric verification system that employs lipreading images as visual features. This verification system uses SVM as a classifier for both subsystems. Principle Component Analysis (PCA) technique is executed for visual features extraction while for the audio feature extraction; Linear Predictive Coding (LPC) technique has been utilized. In this study, we found that the SVM indicator system is able to determine the quality of the speech signal up to 99.66{\%}. At 10 dB SNR, EER performances are observed as 51.13{\%}, 9.3{\%}, and 0.27{\%} for audio only system, fixed weighting system and adaptive weighting system, respectively.",
keywords = "Biometric verification system, reliability estimation, support vector machine",
author = "Ramli, {Dzati Athiar} and {Abdul Samad}, Salina and Aini Hussain",
year = "2010",
doi = "10.1007/978-90-481-8776-8_15",
language = "English",
isbn = "9789048187751",
volume = "60 LNEE",
series = "Lecture Notes in Electrical Engineering",
pages = "165--177",
booktitle = "Lecture Notes in Electrical Engineering",

}

TY - GEN

T1 - An adaptive multibiometric system for uncertain audio condition

AU - Ramli, Dzati Athiar

AU - Abdul Samad, Salina

AU - Hussain, Aini

PY - 2010

Y1 - 2010

N2 - Performances of speaker verification systems are superb in clean noise-free conditions but the reliability of the systems drop severely in noisy environments. In this study, we propose a novel approach by introducing Support Vector Machine (SVM) as indicator system for audio reliability estimation. This approach directly validate the quality of the incoming (claimant) speech signal so as to adaptively change the weighting factor for fusion of both subsystem scores. The effectiveness of this approach has been experimented to a multibiometric verification system that employs lipreading images as visual features. This verification system uses SVM as a classifier for both subsystems. Principle Component Analysis (PCA) technique is executed for visual features extraction while for the audio feature extraction; Linear Predictive Coding (LPC) technique has been utilized. In this study, we found that the SVM indicator system is able to determine the quality of the speech signal up to 99.66%. At 10 dB SNR, EER performances are observed as 51.13%, 9.3%, and 0.27% for audio only system, fixed weighting system and adaptive weighting system, respectively.

AB - Performances of speaker verification systems are superb in clean noise-free conditions but the reliability of the systems drop severely in noisy environments. In this study, we propose a novel approach by introducing Support Vector Machine (SVM) as indicator system for audio reliability estimation. This approach directly validate the quality of the incoming (claimant) speech signal so as to adaptively change the weighting factor for fusion of both subsystem scores. The effectiveness of this approach has been experimented to a multibiometric verification system that employs lipreading images as visual features. This verification system uses SVM as a classifier for both subsystems. Principle Component Analysis (PCA) technique is executed for visual features extraction while for the audio feature extraction; Linear Predictive Coding (LPC) technique has been utilized. In this study, we found that the SVM indicator system is able to determine the quality of the speech signal up to 99.66%. At 10 dB SNR, EER performances are observed as 51.13%, 9.3%, and 0.27% for audio only system, fixed weighting system and adaptive weighting system, respectively.

KW - Biometric verification system

KW - reliability estimation

KW - support vector machine

UR - http://www.scopus.com/inward/record.url?scp=78651539874&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78651539874&partnerID=8YFLogxK

U2 - 10.1007/978-90-481-8776-8_15

DO - 10.1007/978-90-481-8776-8_15

M3 - Conference contribution

AN - SCOPUS:78651539874

SN - 9789048187751

VL - 60 LNEE

T3 - Lecture Notes in Electrical Engineering

SP - 165

EP - 177

BT - Lecture Notes in Electrical Engineering

ER -