Regresar

Identification of normophonic and dysphonic groups in self-organized maps

Abstract:

The voice is a sound that contains characteristics which allow diagnosing various diseases. Studying the disorders that are generated in it is of particular interest in the scientific world. Identify the contributions of voice and automatic strategies in the grouping of people normophonic o dysphonic state of voice es also. This article analyzes the UPM-CTB database of the Center for Biometric Technology of the Polytechnic University of Madrid with 72 characteristics, in which 200 voice samples from 50 patients, male and female gender, both normophonic and dysphonic. As a result it was obtained that it is possible to separate with only two characteristics, patients in normphonic or disphonic groups by gender, using self-organized maps, an unsupervised artificial intelligence technique.