Speaker Recognition Using Wavelet Packet Entropy, I-Vector, and Cosine Distance Scoring

Table 2

The time cost of the different speaker recognition models.

Speaker recognition model	Feature extraction (s/speech)		Speaker classification (s/speech)
	Short vector extraction	I-vector extraction	Training classifier	Recognition
MFCC-GMM	0.46	—	1.92	0.64
MFCC-I-CDS	0.45	2.37	1.81	0.73
WPE-I-CDS	2.81	2.24	1.82	0.75
WPE-I-CDS; (parallel computation)	0.78	2.23	1.82	0.74