System for identifying speaker from given speech signal using MFCC,LPC features and Gaussian Mixture Models
Here we have two systems which are desinged for speaker recognition, One uses the popular MFCC features while the other does the same using Linear Predictive Coeficients.
For MFCC-GMM system, implementation is completely python based. PyaudioAnalysis library has been used for feature extraction of voiced reagion of speaker GMM modeling is done using sklearn package in python
For LPCC-GMM system implementation is partly in Scilab and python For feature extraction Scilab script is used and GMM training is done using sklearn
Various experiments done for hyderparameter optimization can be found in report.