Speaker Recognition System With Pitch Detection Algorithm

Abstract

A simple password is not enough to differentiate between an authorized and a fraudent person. It can be guessed or it can be stolen. So, as technology advances, the security and authentication terms are becoming more intricate and strict, and that has introduced the term ‘biometric’ in our technical dictionary. In this attempt a speaker recognition system has been developed using pitch detection algorithm. The use of the pitch detection algorithm has been done to differentiate between male and female voice. The categorized voice sample then subjected to extract the feature using cepstrum coefficient. Dynamic time warping method is used for the purpose of pattern matching. To evaluate the system each sample utterance of the user is compared with the rest of the utterance in the database. For each comparison, the distance measured is calculated. A lower distance indicates a higher similarity. A lower distance indicates a higher similarity.