Abstract

The usefulness of identifying a person from the characteristics of his voice is increasing. This thesis will describe an approach to speaker identificaion where a neural classifier is used to separate different speakers. Sevel possible solutions will be showed. We will select cepstrum parameters as speaker's feature and explore the ways for solving the problem of speaker identificaiton; the artificial neuarl network is introduces. The random combination of isolated digits from 0 to 9 is specified as identifaction utterances. The system has been evaluated on a database of isalated digit utterance of 20 speakers.