Publication Details
Exploring ANN Back-Ends for i-Vector Based Speaker Age Estimation
Glembek Ondřej, Ing., Ph.D.
Kinnunen Tomi (FIT)
Matějka Pavel, Ing., Ph.D. (DCGM)
age estimation, i-vector, multilayer perceptron
This publication focuses on exploring artificial neural net (ANN) Back-Ends for i-Vector Based Speaker Age Estimation.
We address the problem of speaker age estimation using ivectors. We first compare different i-vector extraction setups and then focus on (shallow) artificial neural net (ANN) backends. We explore ANN architecture, training algorithm and ANN ensembles. The results on NIST 2008 and 2010 SRE data indicate that, after extensive parameter optimization, ANN back-end in combination with i-vectors reaches mean absolute errors (MAEs) of 5.49 (females) and 6.35 (males), which are 4.5% relative improvement in comparison to our support-vector regression (SVR) baseline. Hence, the choice of back-end did not affect the accuracy much; a suggested future direction is therefore focusing more on front-end processing.
@inproceedings{BUT119907,
author="Anna {Silnova} and Ondřej {Glembek} and Tomi {Kinnunen} and Pavel {Matějka}",
title="Exploring ANN Back-Ends for i-Vector Based Speaker Age Estimation",
booktitle="Proceedings of Interspeech 2015",
year="2015",
journal="Proceedings of Interspeech",
volume="2015",
number="09",
pages="3036--3040",
publisher="International Speech Communication Association",
address="Dresden",
isbn="978-1-5108-1790-6",
issn="1990-9772",
url="https://www.fit.vut.cz/research/publication/10971/"
}