A Multiscale Chaotic Feature Extraction Method for Speaker Recognition

Table 1

The combinations of experiment settings.

Experiment group	Speech type	Noise type	Acoustic feature	Nonlinear feature	Chaotic feature (acoustic + nonlinear)
1	Clean speech	—	√
				√
					√
2	Noise speech	White	√
				√
					√
		Babble	√
				√
					√