Text-independent speaker identification system based on the histogram of DCT-cepstrum coefficients

S. Al-Rawahy, A. Hossen, U. Heute

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

There are several known feature sets for text-independent speaker-identification systems, most of which depend on spectral information. Among these feature sets as a most successful one, there is the set of the Mel-Frequency Cepstrum Coefficients (MFCC). This paper introduces a new feature set, namely, the Histogram of the DCT-Cepstrum Coefficients, inspired by the common use of the MFCC, but simpler and faster in computation. A text-independent speaker-identification system based on the DCT-Cepstrum Histogram and Gaussian Mixture Model (GMM) is implemented. The new feature was tested using speech files from the ELSDSR database and TIMIT corpus. The new feature set managed to achieve high efficiency rates with speaker identification accuracy of 100% on 23 speakers from the ELSDSR database, and 99% on 630 speakers from the TIMIT corpus.

Original languageEnglish
Pages (from-to)141-161
Number of pages21
JournalInternational Journal of Knowledge-Based and Intelligent Engineering Systems
Volume16
Issue number3
DOIs
Publication statusPublished - 2012

Fingerprint

Identification (control systems)

Keywords

  • DCT-cepstrum histogram
  • GMM
  • MFCC
  • Speaker identification
  • text-independent

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Control and Systems Engineering

Cite this

@article{164dfacd8fc64a3c90742ce6c02565ee,
title = "Text-independent speaker identification system based on the histogram of DCT-cepstrum coefficients",
abstract = "There are several known feature sets for text-independent speaker-identification systems, most of which depend on spectral information. Among these feature sets as a most successful one, there is the set of the Mel-Frequency Cepstrum Coefficients (MFCC). This paper introduces a new feature set, namely, the Histogram of the DCT-Cepstrum Coefficients, inspired by the common use of the MFCC, but simpler and faster in computation. A text-independent speaker-identification system based on the DCT-Cepstrum Histogram and Gaussian Mixture Model (GMM) is implemented. The new feature was tested using speech files from the ELSDSR database and TIMIT corpus. The new feature set managed to achieve high efficiency rates with speaker identification accuracy of 100{\%} on 23 speakers from the ELSDSR database, and 99{\%} on 630 speakers from the TIMIT corpus.",
keywords = "DCT-cepstrum histogram, GMM, MFCC, Speaker identification, text-independent",
author = "S. Al-Rawahy and A. Hossen and U. Heute",
year = "2012",
doi = "10.3233/KES-2012-0239",
language = "English",
volume = "16",
pages = "141--161",
journal = "International Journal of Knowledge-Based and Intelligent Engineering Systems",
issn = "1327-2314",
publisher = "IOS Press",
number = "3",

}

TY - JOUR

T1 - Text-independent speaker identification system based on the histogram of DCT-cepstrum coefficients

AU - Al-Rawahy, S.

AU - Hossen, A.

AU - Heute, U.

PY - 2012

Y1 - 2012

N2 - There are several known feature sets for text-independent speaker-identification systems, most of which depend on spectral information. Among these feature sets as a most successful one, there is the set of the Mel-Frequency Cepstrum Coefficients (MFCC). This paper introduces a new feature set, namely, the Histogram of the DCT-Cepstrum Coefficients, inspired by the common use of the MFCC, but simpler and faster in computation. A text-independent speaker-identification system based on the DCT-Cepstrum Histogram and Gaussian Mixture Model (GMM) is implemented. The new feature was tested using speech files from the ELSDSR database and TIMIT corpus. The new feature set managed to achieve high efficiency rates with speaker identification accuracy of 100% on 23 speakers from the ELSDSR database, and 99% on 630 speakers from the TIMIT corpus.

AB - There are several known feature sets for text-independent speaker-identification systems, most of which depend on spectral information. Among these feature sets as a most successful one, there is the set of the Mel-Frequency Cepstrum Coefficients (MFCC). This paper introduces a new feature set, namely, the Histogram of the DCT-Cepstrum Coefficients, inspired by the common use of the MFCC, but simpler and faster in computation. A text-independent speaker-identification system based on the DCT-Cepstrum Histogram and Gaussian Mixture Model (GMM) is implemented. The new feature was tested using speech files from the ELSDSR database and TIMIT corpus. The new feature set managed to achieve high efficiency rates with speaker identification accuracy of 100% on 23 speakers from the ELSDSR database, and 99% on 630 speakers from the TIMIT corpus.

KW - DCT-cepstrum histogram

KW - GMM

KW - MFCC

KW - Speaker identification

KW - text-independent

UR - http://www.scopus.com/inward/record.url?scp=84861380344&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84861380344&partnerID=8YFLogxK

U2 - 10.3233/KES-2012-0239

DO - 10.3233/KES-2012-0239

M3 - Article

VL - 16

SP - 141

EP - 161

JO - International Journal of Knowledge-Based and Intelligent Engineering Systems

JF - International Journal of Knowledge-Based and Intelligent Engineering Systems

SN - 1327-2314

IS - 3

ER -