Gaussian mixture model speaker recognition system experiments with CORPORA database
Adam Dąbrowski, Szymon Drgas, Damian Cetnarowicz, Irena Chmielewska · 2007
In this paper current state-of-the-art speaker recognition system was tested with Polish Corpora [4] database. System based on MFCC feature extraction method and Gaussian mixture model were used in several experiments. The minimum amount of data to train a reliable model has been determined. In the second experiment minimum length of a signal to recognize speakers has been determined. In the last part the speaker discriminative properties of Polish phonemes were investigated. The phonemes with the best speaker discriminative properties were determined.