Skip to main content

Table 7 Results obtained from different c-vector configurations on the male part of the NIST SRE 2010 core-extended condition

From: Introducing phonetic information to speaker embedding for speaker verification

Ā 

EER(%)

minDCF08

minDCF10

x-vector

1.96

0.0109

0.3900

c=0.1+ 2-layer sharing

1.12

0.0066

0.2748

c=0.1 + 3-layer sharing

1.21

0.0065

0.2449

c=0.2 + 2-layer sharing

1.18

0.0067

0.3017

c=0.2 + 3-layer sharing

1.24

0.0071

0.2623

c=0.3 + 2-layer sharing

1.29

0.0074

0.2531

c=0.3 + 3-layer sharing

1.46

0.0074

0.2536