Machine learning methods for voice masculinity and femininity scoring and its application to individuals on the autism continuum

Fuling Chen

Research output: ThesisDoctoral Thesis

36 Downloads (Pure)

Abstract

This thesis explores autism's association with voice masculinity-femininity, scrutinizing the Extreme Male Brain and androgyny accounts. It pioneers machine learning-based voice scoring, examining voice-gender links to autistic traits (adults) and autism (children). The developed machine learning models performed well in modelling voice masculinity-femininity. Minor voice gender differences were observed among adults with varying autistic traits. However, autistic children exhibited notably more masculine (boys) and less feminine (girls) voices compared to their non-autistic peers of the same age and sex. The study advances automated voice scoring and delivers insights into voice acoustics, sex differences, masculinity-femininity, and autism continuum associations.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • The University of Western Australia
Supervisors/Advisors
  • Togneri, Roberto, Supervisor
  • Maybery, Murray, Supervisor
  • Tan, Diana, Supervisor
Thesis sponsors
Award date26 Mar 2024
DOIs
Publication statusUnpublished - 2023

Fingerprint

Dive into the research topics of 'Machine learning methods for voice masculinity and femininity scoring and its application to individuals on the autism continuum'. Together they form a unique fingerprint.

Cite this