Automatic gender identification system for Bengali speech

This paper describes a set of features along with different classifiers for gender (male, female) identification. Gender identification is the problem to detect the gender of the speaker of the audio content. The proposed method in this study comprises three stages: i) Corpus construction, iii) Input Feature Vector Construction and iii) Identification using a classifier. In this connection, we have constructed a medium scale Bengali speech corpus from The Bengali Daily Newspaper ProthomAlo.