Skip to content

Publications

Detail of publication

Citation

Zbyněk Zajíc and Marek Hrúz : Fisher Vectors in PLDA Speaker Verification System . The IEEE 13th International Conference on Signal Processing, ICSP 2016, p. 1338-1341, IEEE Press, 2016.

Download PDF

PDF

Abstract

The goal of this paper is to examine the Fisher Vec- tor and incorporate this vector in the PLDA based speaker veri- fication system. The PLDA based system utilizes the Supervector of Statistics extracted from a Gaussian Mixture Model (adopted from the speaker adaptation task) to collect the information about a speaker from a dataset. We compare the efficiency of the PLDA based speaker verification system using Supervector of Statistics and the same system with Fisher vector. The experimental results of these two approaches to the verification task and the fusion of these two systems indicate that the Fisher Vector brings almost the same information to the PLDA verification process as the Supervector of Statistics when sufficient data are available.

Detail of publication

Title: Fisher Vectors in PLDA Speaker Verification System
Author: Zbyněk Zajíc ; Marek Hrúz
Language: English
Date of publication: 6 Nov 2016
Year: 2016
Type of publication: Papers in proceedings of reviewed conferences
Title of journal or book: The IEEE 13th International Conference on Signal Processing, ICSP 2016
Page: 1338 - 1341
ISBN: 978-1-5090-1345-6
Publisher: IEEE Press
/ 2017-01-12 14:35:06 /

Keywords

Speaker Verification, PLDA, Fisher Vector, Su- pervector, iVecto

BibTeX

@INPROCEEDINGS{ZbynekZajic_2016_FisherVectorsin,
 author = {Zbyn\v{e}k Zaj\'{i}c and Marek Hr\'{u}z},
 title = {Fisher Vectors in PLDA Speaker Verification System},
 year = {2016},
 publisher = {IEEE Press},
 journal = {The IEEE 13th International Conference on Signal Processing, ICSP 2016},
 pages = {1338-1341},
 ISBN = {978-1-5090-1345-6},
 url = {http://www.kky.zcu.cz/en/publications/ZbynekZajic_2016_FisherVectorsin},
}