Publication Details
Analysis of BUT Submission in Far-Field Scenarios of VOiCES 2019 Challenge
Plchot Oldřich, Ing., Ph.D. (DCGM)
Zeinali Hossein, Ph.D. (DCGM)
Mošner Ladislav, Ing. (DCGM)
Silnova Anna, M.Sc., Ph.D. (DCGM)
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Novotný Ondřej, Ing., Ph.D.
Glembek Ondřej, Ing., Ph.D.
Far-Field Scenarios, analysis, voices
This paper is a post-evaluation analysis of our efforts inVOiCES 2019 Speaker Recognition challenge. All systems inthe fixed condition are based on x-vectors with different featuresand DNN topologies. The single best system reaches minDCFof 0.38 (5.25% EER) and a fusion of 3 systems yields minDCFof 0.34 (4.87% EER).We also analyze how speaker verification(SV) systems evolved in last few years and show results also onSITW 2016 Challenge. EER on the core-core condition of theSITW 2016 challenge dropped from 5.85% to 1.65% for systemfusions submitted for SITW 2016 and VOiCES 2019, respectively.The less restrictive open condition allowed us to useexternal data for PLDA adaptation and achieve additional smallperformance improvement. In our submission to open condition,we used three x-vector systems and also one system basedon i-vectors.
@inproceedings{BUT159997,
author="Pavel {Matějka} and Oldřich {Plchot} and Hossein {Zeinali} and Ladislav {Mošner} and Anna {Silnova} and Lukáš {Burget} and Ondřej {Novotný} and Ondřej {Glembek}",
title="Analysis of BUT Submission in Far-Field Scenarios of VOiCES 2019 Challenge",
booktitle="Proceedings of Interspeech",
year="2019",
journal="Proceedings of Interspeech",
volume="2019",
number="9",
pages="2448--2452",
publisher="International Speech Communication Association",
address="Graz",
doi="10.21437/Interspeech.2019-2471",
issn="1990-9772",
url="https://www.isca-speech.org/archive/Interspeech_2019/pdfs/2471.pdf"
}