Speaker recognition system based upon classification of Mel-Frequency Cepstral Coefficients (MFCC) using a minimum-distance classifier and a Gaussian Mixture Model (GMM) Log-Likelihood Ratio (LLR) classifier.
The reference speaker recognition system was implemented in MATLAB using training data and test data stored in WAV files. There are tools included in MATLAB and publicly-available libraries to aid in creating this system. For reading in the data sets, we used MATLAB’s wavread function. For feature extraction, we used the melcepst function from Voicebox, a MATLAB toolbox. We used twelve MFCC coefficients (skipping the 0th order coefficient) using 256-sample frames and a 128-sample increment Hamming window. We used custom matching and testing routines based upon minimum-distance classification as described above. For the Gaussian Mixture Models, we used T. N. Vikram’s GMM library, based upon the text Algorithm Collections For Digital Signal Processing Applications Using Matlab by E.S. Gopi.
Please see the project page for details.