Článek ve sborníku konference

SILNOVA Anna, GLEMBEK Ondřej, KINNUNEN Tomi a MATĚJKA Pavel. Exploring ANN Back-Ends for i-Vector Based Speaker Age Estimation. In: Proceedings of Interspeech 2015. Dresden: International Speech Communication Association, 2015, s. 3036-3040. ISBN 978-1-5108-1790-6. ISSN 1990-9772.
Jazyk publikace:angličtina
Název publikace:Exploring ANN Back-Ends for i-Vector Based Speaker Age Estimation
Název (cs):Využití ANN klasifikátorů pro odhad věku řečníka založený na i-vektorech
Strany:3036-3040
Sborník:Proceedings of Interspeech 2015
Konference:INTERSPEECH 2015
Místo vydání:Dresden, DE
Rok:2015
ISBN:978-1-5108-1790-6
Časopis:Proceedings of Interspeech, roč. 2015, č. 09, FR
ISSN:1990-9772
Vydavatel:International Speech Communication Association
URL:http://www.fit.vutbr.cz/research/groups/speech/publi/2015/fedorova_interspeech2015_IS150735.pdf [PDF]
Klíčová slova
age estimation, i-vector, multilayer perceptron
Anotace
Tento článek pojednává o využití artificial neural net (ANN) klasifikátorů pro odhad věku řečníka založený na i-vektorech.
Abstrakt
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.
BibTeX:
@INPROCEEDINGS{
   author = {Anna Silnova and Ond{\v{r}}ej Glembek and Tomi
	Kinnunen and Pavel Mat{\v{e}}jka},
   title = {Exploring ANN Back-Ends for i-Vector Based Speaker
	Age Estimation},
   pages = {3036--3040},
   booktitle = {Proceedings of Interspeech 2015},
   journal = {Proceedings of Interspeech},
   volume = 2015,
 number = 09,
   year = 2015,
   location = {Dresden, DE},
   publisher = {International Speech Communication Association},
   ISBN = {978-1-5108-1790-6},
   ISSN = {1990-9772},
   language = {english},
   url = {http://www.fit.vutbr.cz/research/view_pub.php.cs?id=10971}
}

Vaše IPv4 adresa: 34.239.158.107
Přepnout na https