A real-time application to detect human voice disorders

M. G. Manisha Milani, Murugaiya Ramashini, Murugiah Krishani

20202020 International Conference on Decision Aid Sciences and Application (DASA)12 citationsDOI

Abstract

This paper proposes an automatic method to differentiate healthy and pathological human voices in real-time to construct more accurate medical decisions and allow patients with disorders to seek early medical assistance. The main problem encountered in voice recognition is to identify the appropriate features and a classifier. To address a solution to this problem, this paper introduced a fast and accurate voice recognition method by approaching the MFCC audio processing features with three machine learning algorithms; Decision Tree (DT), Support Vector Machine (SVM), and Artificial Neural Network (ANN). As a result, ANN classifier performed the highest classification accuracy of 87.5%, whilst DT and SVM performed 62.5% and 50%, respectively.

Topics & Concepts

Support vector machineComputer scienceClassifier (UML)Mel-frequency cepstrumArtificial intelligenceDecision treeArtificial neural networkSpeech recognitionMachine learningPattern recognition (psychology)Feature extractionConstruct (python library)Programming languageVoice and Speech DisordersSpeech Recognition and SynthesisMusic and Audio Processing