Table 7 SNR-based relative percent EER reduction rates for i-vector method compared to Drugman’s VAD (second and fourth column) and Rangachari’s method (third and fifth column). Computed by averaging over all noise types
From: A robust polynomial regression-based voice activity detector for speaker verification
SNR level (dB) | Relative EER reduction compared to Drugman’s (male) | Relative EER reduction compared to Rangachari’s (male) | Relative EER reduction compared to Drugman’s (female) | Relative EER reduction compared to Rangachari’s (female) |
|---|---|---|---|---|
− 10 | 24.50 | 28.14 | 19.79 | 30.01 |
− 5 | 32.41 | 31.98 | 25.43 | 40.85 |
0 | 33.16 | 40.13 | 24.29 | 46.99 |
5 | 31.13 | 47.03 | 18.58 | 51.05 |
10 | 28.19 | 47.04 | 16.31 | 56.12 |
Average | 29.87 | 38.86 | 20.88 | 45 |