Study of statistical robust closed set speaker identification with feature and score-based fusion

Musab T. S. Al-Kaltakchi; Wai L. Woo; Satnam S. Dlay; Jonathon A. Chambers

doi:10.1109/SSP.2016.7551807

Study of statistical robust closed set speaker identification with feature and score-based fusion

Musab T. S. Al-Kaltakchi, Wai L. Woo, Satnam S. Dlay, Jonathon A. Chambers

Computer Information Sciences Department

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

12 Citations (Scopus)

Abstract

In this paper, the statistical combination of Power Normalization Cepstral Coefficient (PNCC) and Mel Frequency Cepstral Coefficient (MFCC) features in robust closed set speaker identification is studied. Feature normalization and warping together with late score-based fusion are also exploited to improve performance in the presence of channel and noise effects. In addition, combinations of score and feature-based approaches are considered with early and/or late fusion; these systems use different feature dimensions (16, 32). A 4th order G.712 type IIR filter is employed to represent handset degradation in the channel. Simulation studies based on the TIMIT database confirm the improvement in Speaker Identification Accuracy (SIA) through the combination of PNCC and MFCC features in the presence of handset and Additive White Gaussian Noise (AWGN) effects.

Original language	English
Title of host publication	2016 IEEE Statistical Signal Processing Workshop (SSP)
Publisher	IEEE
ISBN (Electronic)	978-1-4673-7803-1
ISBN (Print)	978-1-4673-7804-8
DOIs	https://doi.org/10.1109/SSP.2016.7551807
Publication status	Published - 25 Aug 2016

Keywords

Robust closed set speaker identification
early and late fusion
handset
AWGN
G.712 handset

Access to Document

10.1109/SSP.2016.7551807

Cite this

@inproceedings{4b162a389447494086b4e7db1d546b28,

title = "Study of statistical robust closed set speaker identification with feature and score-based fusion",

abstract = "In this paper, the statistical combination of Power Normalization Cepstral Coefficient (PNCC) and Mel Frequency Cepstral Coefficient (MFCC) features in robust closed set speaker identification is studied. Feature normalization and warping together with late score-based fusion are also exploited to improve performance in the presence of channel and noise effects. In addition, combinations of score and feature-based approaches are considered with early and/or late fusion; these systems use different feature dimensions (16, 32). A 4th order G.712 type IIR filter is employed to represent handset degradation in the channel. Simulation studies based on the TIMIT database confirm the improvement in Speaker Identification Accuracy (SIA) through the combination of PNCC and MFCC features in the presence of handset and Additive White Gaussian Noise (AWGN) effects.",

keywords = "Robust closed set speaker identification, early and late fusion, handset, AWGN, G.712 handset",

author = "Al-Kaltakchi, {Musab T. S.} and Woo, {Wai L.} and Dlay, {Satnam S.} and Chambers, {Jonathon A.}",

year = "2016",

month = aug,

day = "25",

doi = "10.1109/SSP.2016.7551807",

language = "English",

isbn = "978-1-4673-7804-8",

booktitle = "2016 IEEE Statistical Signal Processing Workshop (SSP)",

publisher = "IEEE",

address = "United States",

}

TY - GEN

T1 - Study of statistical robust closed set speaker identification with feature and score-based fusion

AU - Al-Kaltakchi, Musab T. S.

AU - Woo, Wai L.

AU - Dlay, Satnam S.

AU - Chambers, Jonathon A.

PY - 2016/8/25

Y1 - 2016/8/25

N2 - In this paper, the statistical combination of Power Normalization Cepstral Coefficient (PNCC) and Mel Frequency Cepstral Coefficient (MFCC) features in robust closed set speaker identification is studied. Feature normalization and warping together with late score-based fusion are also exploited to improve performance in the presence of channel and noise effects. In addition, combinations of score and feature-based approaches are considered with early and/or late fusion; these systems use different feature dimensions (16, 32). A 4th order G.712 type IIR filter is employed to represent handset degradation in the channel. Simulation studies based on the TIMIT database confirm the improvement in Speaker Identification Accuracy (SIA) through the combination of PNCC and MFCC features in the presence of handset and Additive White Gaussian Noise (AWGN) effects.

AB - In this paper, the statistical combination of Power Normalization Cepstral Coefficient (PNCC) and Mel Frequency Cepstral Coefficient (MFCC) features in robust closed set speaker identification is studied. Feature normalization and warping together with late score-based fusion are also exploited to improve performance in the presence of channel and noise effects. In addition, combinations of score and feature-based approaches are considered with early and/or late fusion; these systems use different feature dimensions (16, 32). A 4th order G.712 type IIR filter is employed to represent handset degradation in the channel. Simulation studies based on the TIMIT database confirm the improvement in Speaker Identification Accuracy (SIA) through the combination of PNCC and MFCC features in the presence of handset and Additive White Gaussian Noise (AWGN) effects.

KW - Robust closed set speaker identification

KW - early and late fusion

KW - handset

KW - AWGN

KW - G.712 handset

U2 - 10.1109/SSP.2016.7551807

DO - 10.1109/SSP.2016.7551807

M3 - Conference contribution

SN - 978-1-4673-7804-8

BT - 2016 IEEE Statistical Signal Processing Workshop (SSP)

PB - IEEE

ER -

Study of statistical robust closed set speaker identification with feature and score-based fusion

Abstract

Keywords

Access to Document

Fingerprint

Cite this