Cepstral Coefficient-Based Gender Classification Using Audio Signals

Abstract

AbstractIn this paper, audio signal-based gender classification using cepstral coefficients is presented. The audio signals are divided into small frames after the noise removal stage, and Mel-frequency cepstral coefficients and statistical coefficients are extracted to create training and testing data sets. The performance analysis of the system is done by comparing results of different methods. Comparison of the results using MFCC coefficients and statistical coefficients confirms that the proposed method outperforms over the existing classification methods.KeywordsShort-time fourier transform (STFT)Mel-frequency cepstral coefficients (MFCC)Wavelet denoiserStatistical coefficientsSupport vector machine (SVM)