A Review: Voice Pathology Classification Using Machine Learning

Abstract

Voice pathology detection requires the presence of a specialist doctor and time to treat each patient, but it is not always possible to have a doctor who can treat all patients at once and at one precise time. For residents of remote areas, it is all expensive equipment that must be provided. Or even for people who may not be aware of having any voice pathology. Our goal is to design a diagnostic aid system to detect whether the voice is pathological or healthy, so that the patient can be referred to a doctor or not without being moved from the start. Our system is based on the classification, by SVM "Support Vector Machine", using MFCCs "Mel Frequency Cepstral Coefficients" extracted from the patient's voice. The learning and testing of our system are done using the SVD database "Saarbruecken Voice Database"