Gender identification from speech signal by examining the speech production characteristics

The term gender identification deals with finding out the gender of a person from his or her voice. Gender identification has been implemented in several Automatic Speaker Recognition (ASR) systems and has proved to be of great significance. The use of gender identification in today's technology makes it easier for user authentication and identification in high security systems. In this paper, we have discussed about the gender identification process for speech signals using three different features namely Pitch using autocorrelation, Signal energy and Mel Frequency Cepstral Coefficients (MFCCs). A linear Support Vector Machine (SVM) classifier was used for classification of features extracted from the speech signal using signal processing methods. Two sets of experiments were performed - in the first experiment, one speech file was tested against one training file as a one-on-one experiment. In the second experiment, one speech file was tested against three training files. The average accuracy of the second experiment was slightly higher than the first experiment. Performance evaluation results are encouraging. The approach can be used in wide range of applications.