Efficient speaker recognition using approximated cross entropy (ACE)

Hagai Aronowitz; David Burshtein

doi:10.1109/TASL.2007.902059

Efficient speaker recognition using approximated cross entropy (ACE)

Hagai Aronowitz^*, David Burshtein

^*Corresponding author for this work

School of Electrical Engineering

Research output: Contribution to journal › Article › peer-review

31 Scopus citations

Abstract

Techniques for efficient speaker recognition are presented. These techniques are based on approximating Gaussian mixture modeling (GMM) likelihood scoring using approximated cross entropy (ACE). Gaussian mixture modeling is used for representing both training and test sessions and is shown to perform speaker recognition and retrieval extremely efficiently without any notable degradation in accuracy compared to classic GMM-based recognition. In addition, a GMM compression algorithm is presented. This algorithm decreases considerably the storage needed for speaker retrieval.

Original language	English
Article number	4291589
Pages (from-to)	2033-2043
Number of pages	11
Journal	IEEE Transactions on Audio, Speech and Language Processing
Volume	15
Issue number	7
DOIs	https://doi.org/10.1109/TASL.2007.902059
State	Published - Sep 2007

Keywords

Speaker identification
Speaker indexing
Speaker recognition
Speaker retrieval
Speaker verification

Access to Document

10.1109/TASL.2007.902059

Cite this

@article{61929668ef834d1aab3a54e0681a692d,

title = "Efficient speaker recognition using approximated cross entropy (ACE)",

abstract = "Techniques for efficient speaker recognition are presented. These techniques are based on approximating Gaussian mixture modeling (GMM) likelihood scoring using approximated cross entropy (ACE). Gaussian mixture modeling is used for representing both training and test sessions and is shown to perform speaker recognition and retrieval extremely efficiently without any notable degradation in accuracy compared to classic GMM-based recognition. In addition, a GMM compression algorithm is presented. This algorithm decreases considerably the storage needed for speaker retrieval.",

keywords = "Speaker identification, Speaker indexing, Speaker recognition, Speaker retrieval, Speaker verification",

author = "Hagai Aronowitz and David Burshtein",

note = "Funding Information: Manuscript received February 15, 2007; revised May 15, 2007. This work was supported in part by Muscle, a European network of excellence funded by the EC 6th Framework IST Program. The associate editor coordinating the review of this manuscript and approving it for publication was Dr. Jean-Fran{\c c}ois Bonastre.",

year = "2007",

month = sep,

doi = "10.1109/TASL.2007.902059",

language = "אנגלית",

volume = "15",

pages = "2033--2043",

journal = "IEEE Transactions on Audio, Speech and Language Processing",

issn = "1558-7916",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "7",

}

TY - JOUR

T1 - Efficient speaker recognition using approximated cross entropy (ACE)

AU - Aronowitz, Hagai

AU - Burshtein, David

N1 - Funding Information: Manuscript received February 15, 2007; revised May 15, 2007. This work was supported in part by Muscle, a European network of excellence funded by the EC 6th Framework IST Program. The associate editor coordinating the review of this manuscript and approving it for publication was Dr. Jean-François Bonastre.

PY - 2007/9

Y1 - 2007/9

N2 - Techniques for efficient speaker recognition are presented. These techniques are based on approximating Gaussian mixture modeling (GMM) likelihood scoring using approximated cross entropy (ACE). Gaussian mixture modeling is used for representing both training and test sessions and is shown to perform speaker recognition and retrieval extremely efficiently without any notable degradation in accuracy compared to classic GMM-based recognition. In addition, a GMM compression algorithm is presented. This algorithm decreases considerably the storage needed for speaker retrieval.

AB - Techniques for efficient speaker recognition are presented. These techniques are based on approximating Gaussian mixture modeling (GMM) likelihood scoring using approximated cross entropy (ACE). Gaussian mixture modeling is used for representing both training and test sessions and is shown to perform speaker recognition and retrieval extremely efficiently without any notable degradation in accuracy compared to classic GMM-based recognition. In addition, a GMM compression algorithm is presented. This algorithm decreases considerably the storage needed for speaker retrieval.

KW - Speaker identification

KW - Speaker indexing

KW - Speaker recognition

KW - Speaker retrieval

KW - Speaker verification

UR - http://www.scopus.com/inward/record.url?scp=64249095146&partnerID=8YFLogxK

U2 - 10.1109/TASL.2007.902059

DO - 10.1109/TASL.2007.902059

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

AN - SCOPUS:64249095146

SN - 1558-7916

VL - 15

SP - 2033

EP - 2043

JO - IEEE Transactions on Audio, Speech and Language Processing

JF - IEEE Transactions on Audio, Speech and Language Processing

IS - 7

M1 - 4291589

ER -

Efficient speaker recognition using approximated cross entropy (ACE)

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this