Table 3 Average ASV performance (Miss-10 metric, %) as a function of noise type and condition
From: Biomimetic multi-resolution analysis for robust speaker recognition
Cond. | Babble (0–24 dB) | White (0–24 dB) | Subway (0–24 dB) | Reverb (200–1,200 ms) | ||||
|---|---|---|---|---|---|---|---|---|
MFCC | Cortical | MFCC | Cortical | MFCC | Cortical | MFCC | Cortical | |
1 | 5.3 | 3.2 | 17.2 | 7.9 | 9.4 | 6.0 | 9.1 | 4.6 |
2 | 16.6 | 10.8 | 31.3 | 18.1 | 22.9 | 16.4 | 19.8 | 9.5 |
3 | 6.2 | 4.4 | 14.0 | 9.0 | 9.2 | 6.9 | 26.5 | 21.9 |
4 | 8.1 | 5.7 | 22.9 | 13.7 | 13.0 | 9.9 | 16.9 | 9.9 |
5 | 4.9 | 3.7 | 9.4 | 6.7 | 7.0 | 5.3 | 20.3 | 11.8 |
6 | 11.6 | 10.2 | 13.6 | 11.4 | 13.2 | 11.2 | 35.6 | 25.1 |
7 | 11.8 | 12.5 | 18.6 | 16.6 | 13.0 | 16.5 | 34.2 | 26.6 |
8 | 3.2 | 2.9 | 7.7 | 5.0 | 4.7 | 3.9 | 17.0 | 9.4 |
9 | 6.5 | 5.5 | 21.4 | 15.4 | 11.0 | 9.9 | 9.0 | 4.6 |