Text Dependent Speaker Recognition Study
Hui-Hong Xu · 2015
The speaker recognition is a sort of biometrics according to person's sound.This paper proposed a method that extracted characteristic parameter from sound signal by LPCC and MFCC.Improving LBG algorithm, training and testing the samples by continuous left-right HMM, A speaker recognition algorithm was given.Trough experiment, the result in 4 changes continuous left-right HMM is best.