A person identification method in CUG using voice pitch analysis

Abstract

Acoustic features of voice are different according to gender due to physiological differences of the male and female speakers. Since this difference is observed in frequency range of voice, analyzing voice pitch related to this feature is useful in person identification. In this paper, we first propose a person identification method by measuring maximum and minimum frequencies of person's voice. We then propose a speech recognition method, which receives the person's voice via microphone, and transforms his/her voice to text. The proposed methods is used in a meeting management system, where it is useful to identify a person in a closed user group (CUG), and to keep a track of his/her speech. © 2014 IEEE.