Using regularized fisher discriminant analysis to improve the performance of Gaussian supervector in session and device identification

Hung Fat Frank Leung; Yuechi Jiang

doi:10.1109/IJCNN.2017.7965921

Using regularized fisher discriminant analysis to improve the performance of Gaussian supervector in session and device identification

Hung Fat Frank Leung, Yuechi Jiang

The Hong Kong Polytechnic University

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

8 Citations (Scopus)

Abstract

In this paper, we propose Regularized Fisher Discriminant Analysis (RFDA) as a projection method applied on Gaussian Supervector (GSV). GSV was originally applied on speaker recognition and verification, and has exhibited good performance. Recently GSV has also been applied in audio forensics area, such as recording device identification. It has been shown that GSV can also capture useful information related to the recording device. In this paper, we show that GSV can also be applied in telephone session identification. However, although GSV can capture useful information for different identification purposes, the performance of the raw GSV may not be so good. Thus, we apply RFDA-based projection method on the raw GSV, and find that this projection method can significantly improve the performance of the raw GSV, in both telephone session identification and recording device identification tasks.

Original language	English
Title of host publication	2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings
Publisher	IEEE
Pages	705-712
Number of pages	8
Volume	2017-May
ISBN (Electronic)	9781509061815
DOIs	https://doi.org/10.1109/IJCNN.2017.7965921
Publication status	Published - 30 Jun 2017
Event	2017 International Joint Conference on Neural Networks, IJCNN 2017 - Anchorage, United States Duration: 14 May 2017 → 19 May 2017

Conference

Conference	2017 International Joint Conference on Neural Networks, IJCNN 2017
Country/Territory	United States
City	Anchorage
Period	14/05/17 → 19/05/17

Keywords

Audio forensics
Projected Gaussian Supervector
Recording device identification
Regularized Fisher Discriminant Analysis
Telephone session identification

ASJC Scopus subject areas

Software
Artificial Intelligence

More information

10.1109/IJCNN.2017.7965921

http://hdl.handle.net/10397/107242

Cite this

@inproceedings{46b118f821ef457188a3ffa65f51570c,

title = "Using regularized fisher discriminant analysis to improve the performance of Gaussian supervector in session and device identification",

abstract = "In this paper, we propose Regularized Fisher Discriminant Analysis (RFDA) as a projection method applied on Gaussian Supervector (GSV). GSV was originally applied on speaker recognition and verification, and has exhibited good performance. Recently GSV has also been applied in audio forensics area, such as recording device identification. It has been shown that GSV can also capture useful information related to the recording device. In this paper, we show that GSV can also be applied in telephone session identification. However, although GSV can capture useful information for different identification purposes, the performance of the raw GSV may not be so good. Thus, we apply RFDA-based projection method on the raw GSV, and find that this projection method can significantly improve the performance of the raw GSV, in both telephone session identification and recording device identification tasks.",

keywords = "Audio forensics, Projected Gaussian Supervector, Recording device identification, Regularized Fisher Discriminant Analysis, Telephone session identification",

author = "Leung, {Hung Fat Frank} and Yuechi Jiang",

year = "2017",

month = jun,

day = "30",

doi = "10.1109/IJCNN.2017.7965921",

language = "English",

volume = "2017-May",

pages = "705--712",

booktitle = "2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings",

publisher = "IEEE",

address = "United States",

note = "2017 International Joint Conference on Neural Networks, IJCNN 2017 ; Conference date: 14-05-2017 Through 19-05-2017",

}

Leung, HFF & Jiang, Y 2017, Using regularized fisher discriminant analysis to improve the performance of Gaussian supervector in session and device identification. in 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. vol. 2017-May, 7965921, IEEE, pp. 705-712, 2017 International Joint Conference on Neural Networks, IJCNN 2017, Anchorage, United States, 14/05/17. https://doi.org/10.1109/IJCNN.2017.7965921

Using regularized fisher discriminant analysis to improve the performance of Gaussian supervector in session and device identification. / Leung, Hung Fat Frank; Jiang, Yuechi.
2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. Vol. 2017-May IEEE, 2017. p. 705-712 7965921.

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Using regularized fisher discriminant analysis to improve the performance of Gaussian supervector in session and device identification

AU - Leung, Hung Fat Frank

AU - Jiang, Yuechi

PY - 2017/6/30

Y1 - 2017/6/30

N2 - In this paper, we propose Regularized Fisher Discriminant Analysis (RFDA) as a projection method applied on Gaussian Supervector (GSV). GSV was originally applied on speaker recognition and verification, and has exhibited good performance. Recently GSV has also been applied in audio forensics area, such as recording device identification. It has been shown that GSV can also capture useful information related to the recording device. In this paper, we show that GSV can also be applied in telephone session identification. However, although GSV can capture useful information for different identification purposes, the performance of the raw GSV may not be so good. Thus, we apply RFDA-based projection method on the raw GSV, and find that this projection method can significantly improve the performance of the raw GSV, in both telephone session identification and recording device identification tasks.

AB - In this paper, we propose Regularized Fisher Discriminant Analysis (RFDA) as a projection method applied on Gaussian Supervector (GSV). GSV was originally applied on speaker recognition and verification, and has exhibited good performance. Recently GSV has also been applied in audio forensics area, such as recording device identification. It has been shown that GSV can also capture useful information related to the recording device. In this paper, we show that GSV can also be applied in telephone session identification. However, although GSV can capture useful information for different identification purposes, the performance of the raw GSV may not be so good. Thus, we apply RFDA-based projection method on the raw GSV, and find that this projection method can significantly improve the performance of the raw GSV, in both telephone session identification and recording device identification tasks.

KW - Audio forensics

KW - Projected Gaussian Supervector

KW - Recording device identification

KW - Regularized Fisher Discriminant Analysis

KW - Telephone session identification

UR - http://www.scopus.com/inward/record.url?scp=85031041271&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2017.7965921

DO - 10.1109/IJCNN.2017.7965921

M3 - Conference article published in proceeding or book

VL - 2017-May

SP - 705

EP - 712

BT - 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings

PB - IEEE

T2 - 2017 International Joint Conference on Neural Networks, IJCNN 2017

Y2 - 14 May 2017 through 19 May 2017

ER -

Using regularized fisher discriminant analysis to improve the performance of Gaussian supervector in session and device identification

Abstract

Conference

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this