A Multiscale Chaotic Feature Extraction Method for Speaker Recognition

Table 5

The EER value under babble noise speech condition.

SNR (dB)	Acoustic feature	Nonlinear feature	Chaotic feature	i-vector model
0	27.589	30.142	20.092	26.172
5	23.897	25.563	13.095	16.077
10	20.481	21.792	11.607	16.981
15	12.074	13.346	8.103	10.252
20	6.547	5.897	2.918	4.266
25	2.801	2.105	1.691	2.678
30	1.822	2.185	0.901	1.587
Average	13.60	14.43	8.34	11.14