EURASIP Journal on Audio, Speech, and Music Processing

Table 2 The contribution rate (%) of every principle component

From: A sub-band-based feature reconstruction approach for robust speaker recognition

	FB	SB1	SB2
1	50.161	69.455	66.811
2	22.291	13.475	17.025
3	6.935	3.292	6.812
4	5.236	2.315	3.084
5	3.424	2.315	2.051
6	1.945	2.139	1.345
7	1.708	1.514	0.951
8	1.218	1.185	0.715
9	1.125	0.898	0.490
10	1.002	0.651	0.327
11	0.869	0.445	0.233
12	0.698	0.240	0.157

Back to article page