Text-independent speaker recognition based on vowel spotting

An automatic text-independent speaker recognition system is presented, which is suitable for identification as well as for verification purposes. The system is based on spotting the stable part of the vowel phonemes of the test utterances, extracting parameter vectors and classifying them to a speaker-dependent vowel reference database. The system was tested over a period of four months with a population of 12 male and female speakers with non-correlated training and test data. The accuracy of the system as measured by experimentation is satisfactory considering that the training utterances per speaker do not exceed 50 sec and the test utterances 1 sec in average