Intra-frame cepstral sub-band weighting and histogram equalization for noise-robust speech recognition

EURASIP Journal on Audio, Speech, and Music Processing

Table 3 The recognition accuracy results (%) of the MFCC baseline, CHN, and WS-HEQ with structure II

Method		Set A	Set B	Set C	Average	RR
MFCC		59.24	56.37	67.53	59.75	-
CHN		79.28	81.53	79.98	80.32	51.11
${WS-HEQ}_{II}^{(1)}$	α=1.0	81.75	84.61	80.81	82.70	57.03
	α=0.6	84.13	86.16	84.39	84.99	62.71
${WS-HEQ}_{II}^{(2)}$	α=1.0	82.59	84.93	82.03	83.41	58.79
	α=0.6	83.54	85.75	83.83	84.48	61.44
${WS-HEQ}_{II}^{(3)}$	α=1.0	80.55	83.99	79.80	81.77	54.72
	α=0.7	83.25	85.10	83.50	84.04	60.35
${WS-HEQ}_{II}^{(4)}$	α=1.0	80.83	83.44	80.37	81.78	54.73
	α=0.6	82.30	83.45	82.90	82.88	57.47

They are for different test sets while averaged over five SNR conditions (20 to 0 dB) as to the Aurora-2 database. RR (%) is the relative error rate reduction compared with the MFCC baseline.