Detail výsledku
DNN-based SRE Systems in Multi-Language Conditions
Matějka Pavel, Ing., Ph.D., UPGM (FIT)
Glembek Ondřej, Ing., Ph.D., UPGM (FIT)
Plchot Oldřich, Ing., Ph.D., UPGM (FIT)
Grézl František, Ing., Ph.D., UPGM (FIT)
Burget Lukáš, doc. Ing., Ph.D., UPGM (FIT)
Černocký Jan, prof. Dr. Ing., UPGM (FIT)
This work studies the usage of the (currently state-of-the-art) Deep NeuralNetworks (DNN) i-vector/PLDA-based speaker recognition systems inmulti-language (especially non-English) conditions. On the ``Language Pack''of the PRISM set, we evaluate the systems' performance using NIST's standardmetrics. We study the use of multi-lingual DNN in place of the originalEnglish DNN on these multi-language conditions. We show that not only the gainfrom using DNNs vanishes, but also the DNN-based systems tend to producede-calibrated scores under the studied conditions. This work gives suggestionsfor directions of future research rather than any particular solutions.
speaker recognition, multilinguality, DNN
@techreport{BUT168427,
author="Ondřej {Novotný} and Pavel {Matějka} and Ondřej {Glembek} and Oldřich {Plchot} and František {Grézl} and Lukáš {Burget} and Jan {Černocký}",
title="DNN-based SRE Systems in Multi-Language Conditions",
year="2016",
publisher="Faculty of Information Technology BUT",
address="Brno",
pages="5",
url="http://www.fit.vutbr.cz/research/groups/speech/publi/2016/dnn-based-sre_TECH_REP_v0.pdf"
}