Auditory Sparse Representation for Robust Speaker Recognition Based on Tensor Structure

EURASIP Journal on Audio, Speech, and Music Processing

Table 4 Identification accuracy in four noisy conditions (subway, car noise, babble, and exhibition hall) for Aurora2 noise testing dataset.

(%)	SNR	ANTCC	MFCC	LPCC	RASTA-PLP
Subway	5 dB	26.36	2.73	5.45	14.55
	10 dB	63.64	16.36	11.82	39.09
	15 dB	75.45	44.55	34.55	57.27
	20 dB	89.09	76.36	60.0	76.36
Babble	5 dB	43.27	16.36	15.45	22.73
	10 dB	62.73	51.82	33.64	57.27
	15 dB	78.18	79.09	66.36	86.36
	20 dB	87.27	93.64	86.36	92.73
Car noise	5 dB	19.09	5.45	3.64	8.18
	10 dB	30.91	17.27	10.91	35.45
	15 dB	60.91	44.55	33.64	60.91
	20 dB	78.18	78.18	59.09	79.45
Exhibition hall	5 dB	24.55	1.82	2.73	13.64
	10 dB	62.73	20.0	19.09	31.82
	15 dB	85.45	50.0	44.55	59.09
	20 dB	95.45	76.36	74.55	82.73