This thesis investigates computational tools and software applications supporting the evaluation, habilitation, and rehabilitation of human auditory perception. Virtual reality and machine learning techniques ported to mobile devices add ecological validity to hearing tests and improve their potential to capture diverse quantitative data with no additional effort from participants. Acoustic localization is set in a virtual environment with real-time kinematic tracking to analyze the role of spontaneous head movements in normal-hearing listeners. An experimental procedure designed in this setting made it possible to describe, from this point of view, compensatory strategies in children with hearing aids and cochlear implants under various hearing conditions. A web application is presented, employing formal tools such as Gaussian process classification, to efficiently determine the individual’s just-noticeable difference in interaural time, and facilitate the sharing of results between research centers. The last application presented involves a psychoacoustic procedure for measuring frequency discrimination ability with a Bayesian active learning approach to personalize, speed up, and improve the robustness of the test session, and with game dynamics to increase participant involvement. Using mobile devices enables the relocation of assessments from clinical settings to the home environment, allowing participants to engage in training remotely. Results confirm significant improvements in pitch discrimination thresholds among children with unilateral cochlear implants, highlighting the potential for these innovative approaches in hearing research and rehabilitation. Overall, each application improves the understanding of the individual auditory perception and reduces the test session time with the same accuracy as other established methods. These evaluations can be instrumental in developing tailor-made rehabilitation and training programs, and they can be useful for customizing virtual environments and improving their accessibility.
Applications of Computer Interactions and Intelligence to Auditory Tests
GULLI, ANDREA
2024
Abstract
This thesis investigates computational tools and software applications supporting the evaluation, habilitation, and rehabilitation of human auditory perception. Virtual reality and machine learning techniques ported to mobile devices add ecological validity to hearing tests and improve their potential to capture diverse quantitative data with no additional effort from participants. Acoustic localization is set in a virtual environment with real-time kinematic tracking to analyze the role of spontaneous head movements in normal-hearing listeners. An experimental procedure designed in this setting made it possible to describe, from this point of view, compensatory strategies in children with hearing aids and cochlear implants under various hearing conditions. A web application is presented, employing formal tools such as Gaussian process classification, to efficiently determine the individual’s just-noticeable difference in interaural time, and facilitate the sharing of results between research centers. The last application presented involves a psychoacoustic procedure for measuring frequency discrimination ability with a Bayesian active learning approach to personalize, speed up, and improve the robustness of the test session, and with game dynamics to increase participant involvement. Using mobile devices enables the relocation of assessments from clinical settings to the home environment, allowing participants to engage in training remotely. Results confirm significant improvements in pitch discrimination thresholds among children with unilateral cochlear implants, highlighting the potential for these innovative approaches in hearing research and rehabilitation. Overall, each application improves the understanding of the individual auditory perception and reduces the test session time with the same accuracy as other established methods. These evaluations can be instrumental in developing tailor-made rehabilitation and training programs, and they can be useful for customizing virtual environments and improving their accessibility.File | Dimensione | Formato | |
---|---|---|---|
Computer Interactions and Intelligence in Ecological Auditory Tests for Children with Hearing Losses.pdf
embargo fino al 09/03/2026
Dimensione
20.42 MB
Formato
Adobe PDF
|
20.42 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/164565
URN:NBN:IT:UNIUD-164565