A Multiscale Chaotic Feature Extraction Method for Speaker Recognition

Table 4

The EER value under white noise speech condition.

SNR (dB)	Acoustic feature	Nonlinear feature	Chaotic feature	i-vector model
30	2.156	2.785	1.59	1.963
25	3.167	3.389	2.531	3.144
20	5.148	5.987	4.121	4.907
15	9.746	10.182	7.115	9.755
10	17.929	18.114	14.152	18.122
5	29.535	30.157	25.144	32.696
0	37.005	39.142	20.504	33.134
Average	14.96	15.68	10.74	14.82