Tomaszewska, Julia Z. and Georgakis, Apostolos (2023) Electroglottography in Medical Diagnostics of Vocal Tract Pathologies: A Systematic Review. Journal of Voice.

Preview

PDF 1-s2.0-S0892199723003880-main.pdf - Published Version Available under License Creative Commons Attribution. Download (1MB) | Preview

Abstract

Electroglottography (EGG) is a technology developed for measuring the vocal fold contact area
during human voice production. Although considered subjective and unreliable as a sole diagnostic method,
with the correct application of relevant computational methods, it can constitute a most promising non-invasive
voice disorder diagnostic tools in a form of a digital vocal tract pathology classifier. The aim of the following
study is to gather and evaluate currently existing digital voice quality assessment systems and vocal tract abnormality classification systems that rely on the use of electroglottographic bio-impedance signals. To fully
comprehend the findings of this review, first the subject of EGG is introduced. For that, we summarise most
relevant existing research on EGG with a particular focus on its application in diagnostics. Then, we move on to
the focal point of this work, which is describing and comparing the existing EGG-based digital voice pathology
classification systems. With the application of PRISMA model, 13 articles were chosen and analysed in detail.
Direct comparison between chosen studies brought us to pivotal conclusions, which have been described in
Section 5 of this report. Meanwhile, certain limitations arising from the literature were identified, such as
questionable understanding of the nature of EGG bio-impedance signals. The appropriate recommendations
for future work were made, including the application of different methods for EGG feature extraction, as well
as the need for continuous EGG datasets development containing signals gathered in various conditions and
with different equipments.

Item Type:	Article
Identifier:	10.1016/j.jvoice.2023.12.004
Subjects:	Computing
Depositing User:	Users 627 not found.
Date Deposited:	18 Mar 2024 13:33
Last Modified:	04 Nov 2024 11:16
URI:	https://repository.uwl.ac.uk/id/eprint/11319

Downloads

Actions (login required)

View Item