The Formant-Emphasized Feature Vector for Speech Recognition in Noisy Condition

Mel-frequency cepstral coefficients are widely used as the feature for speech recognition. In MFCC extraction process, the spectrum, obtained by Fourier transform of input speech signal is divided by mel-frequency bands, and each ban energy is extracted for the each frequency band. The coefficients are extracted by the discrete cosine transform of the obtained band energy. In this paper, we calculate the output energy for each bandpass filter by taking the weighting function when applying mel-frequency scaled bandpass filter. The weighting function is Gaussian distributed function whose center is at the formant frequency. In the experiments, we can see the comparative performance with the standard MFCC in clean condition, and the better performance in worse condition using method we proposed.