A new MFCC improvement method for robust ASR
Hojatollah Yeganeh, Seyed Mohammad Ahadi, Ali Ziaei · 2008
The Mel-frequency cepstral coefficients (MFCC) are widely used for speech recognition. However, MFCC-based speech recognition performance degrades in presence of additive noise. In this paper, we propose a set of noise-robust features based on conventional MFCC feature extraction method. Our proposed method consists of two steps. In the first step, Mel sub-band spectral subtraction is carried out. The second step consists of estimating SNR in each sub-band and defining a weight parameter based on this estimation. The weighting has been carried out in a way that gives more important roles in cepstrum parameter formation to sub-bands that are less affected by noise. Experimental results indicate that this method achieves improved performance for ASR in noisy environments. Furthermore, due to the simplicity of the implementation of our method, its computational overhead relative to MFCC is quite small.