Text-indepentent speaker recognition with polynomial classifiers
Aarti Bakshi · 2010
This paper concerns the design and implementation of a real-time text independent speaker recognition system. This involves speaker identification by determining who the speaker is. The system utilizes Polynomial classifiers to match the templates and for feature extraction mel-frequency cepstral coefficients (MFCC) are used. Before extracting the features the speech signal is preprocessed using Pre-emphasis Framing & Windowing. There after twelve mel-cepstral coefficients are derived from each frame. In polynomial classifiers, training the classifiers is done for each speaker model with discriminative training with a mean squared error criterion.