METHOD AND APPARATUS FOR VOICE-BASED GENDER IDENTIFICATION

摘要

A gender recognizing method based on a audio signal and an apparatus thereof are provided to extract a feature vector in which a voiceless sound is removed and compare the feature vector with a Gaussian mixture model and a voiced sound, thereby increasing accuracy of recognizing gender. A Gaussian mixture model in which a voiceless sound is removed with regard to a plurality of audio signals is generated(S100). An audio signal of a speaker is received(S200). A feature vector in which a voiceless sound is removed with regard to the audio signal is extracted(S300). The gender of the audio signal is determined by comparing the extracted feature vector with the Gaussian mixture model.