Skip to main content

Table 5 Effects of semi-fragile watermarking on speaker verification performance for different speech databases

From: Semi-fragile digital speech watermarking for online speaker recognition

Database

System

EER (%) without fragile WM

EER (%) with fragile WM

TIMIT

i-vector + MFCC

0.71

0.74

i-vector + LPRC

1.45

1.46

GMM-UBM + MFCC

0.79

0.80

GMM-UBM + LPRC

1.90

1.90

MIT

i-vector + MFCC

15.04

15.17

i-vector + LPRC

24.20

24.26

GMM-UBM + MFCC

46

46.02

GMM-UBM + LPRC

49.43

49.93

MOBIO

i-vector + MFCC

45.96

45.96

i-vector + LPRC

46

46

GMM-UBM + MFCC

46.66

46.66

GMM-UBM + LPRC

47

47