Classification of voice aging using ANN and glottal signal parameters

Abstract

Classification of voice aging has many applications in health care and geriatrics. This work focuses on finding the most relevant parameters to classify voice aging. The most significant parameters extracted from the glottal signal are chosen to identify the voice aging process of men and women. After analyzing their statistics, the chosen parameters are used as entries to a neural network to classify male and female Brazilian speakers in three different age groups: young (from 15 to 30 years old), adult (from 31 to 60 years old), and senior (from 61 to 90 years old). The corpus used for this work was composed by one hundred and twenty Brazilian speakers (both males and females) of different ages. As compared to similar works, we employ a larger corpus and obtain a superior classification rate.