Identificação de patologias da laringe através do sinal da voz

Abstract

Aiming to study a solution to the adversities of current diagnostic methods, this monography sought to identify and diagnose laryngeal pathologies in a non-invasive way. The research was developed through the extraction of acoustic measures – such as pitch, jitter, shimmer and others – and subsequent application in machine learning methods, in this case the Support Vector Machines (SVM) and Artificial Neural Networks (ANN) classifiers. Based on the statistical data of the models tested, analyses of the effectiveness and the possibility of application in medical diagnoses were carried out.