Gender Classification with Voice Intensity

Abstract

In this paper, a new approach to classify the two types of gender, male and female, is proposed and implemented. The research is based on the idea that male and female can be classified on the basis of the intensity of their utterances. To implement the idea, the Simpson's rule of numerical integration is used to calculate the voice intensity by measuring area under the normalized curve. The curve is obtained after multiplying a factor with the polynomial of degree three (fitted through the peaks), where peaks are found from each frame of speech utterance when it is divided into frames of length of 20 milliseconds. A certain level of amplitude is considered while finding the peaks. To make the decision about the type of gender an adaptively adjusted threshold is used. If the area of an utterance is above the threshold the gender type is a female otherwise male. Experiments show that result is promising that is 96.44%.