Phoniatric System Based on Acoustical Analysis for Early Detection of Anomalies in Voice Production

Ana Laura Cazarin, Eladio Cardiel, Laura I. Garay-Jimenez, Pablo Rogelio Hernandez, Victor Manuel Valadez Jimenez

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

An objective evaluation of voice quality using phoniatric parameters to analyze acoustic signals generated during phonations has been developed. The system used the Voice Handicap Index questionnaire to quantify the perception of the vocal capacity of people, complemented with the measurement of voice acoustic signal parameters obtained from sustained vocalizations. The parameters were statistically evaluated, and then a systematic analysis of seven classifiers with the preselected parameter was done with and without PCA, KNN, BPNN, and SVM obtained the best performance with a mean accuracy of 97.4%, 96%, and 94.7%. The proposal can be used to determine alterations in the production of the voice for medical diagnostics.

Original languageEnglish
Title of host publicationCCE 2021 - 2021 18th International Conference on Electrical Engineering, Computing Science and Automatic Control
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665400299
DOIs
StatePublished - 2021
Event18th International Conference on Electrical Engineering, Computing Science and Automatic Control, CCE 2021 - Mexico City, Mexico
Duration: 10 Nov 202112 Nov 2021

Publication series

NameCCE 2021 - 2021 18th International Conference on Electrical Engineering, Computing Science and Automatic Control

Conference

Conference18th International Conference on Electrical Engineering, Computing Science and Automatic Control, CCE 2021
Country/TerritoryMexico
CityMexico City
Period10/11/2112/11/21

Keywords

  • Acoustical Analysis
  • Phoniatric
  • Voice Quality

Fingerprint

Dive into the research topics of 'Phoniatric System Based on Acoustical Analysis for Early Detection of Anomalies in Voice Production'. Together they form a unique fingerprint.

Cite this