A frequency warping approach for vocal tract length normalization

A method of vocal tract length normalization (VTLN) is proposed. It uses bilinear transform (BLT) to modify the filterbank in Mel-frequency cepstrum based on the average third formant F₃. The effectiveness of this method is examined on vowel and isolated digit recognitions. The baseline vowel recognition models are trained on males data and the baseline isolated digit models are trained on adult men´s data respectively. When the MFCC coefficients of test data are transformed by BLT, the recognition accuracy of females´ vowels is improved by 11.67% and the recognition accuracies of adult women and children´s isolated digits are improved by 19.5% and 13% respectively.