Enrico Bertolami
Adaptive Mode Decomposition Algorithms: Comparative Analysis and Application to Speech Signals.
Rel. Cecilia Surace, Marco Civera. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Biomedica, 2026
Abstract
Adaptive mode decomposition algorithms are methods used to decompose a signal into a sum of amplitude- and frequency-modulated components. This allows for the study of nonstationary and nonlinear signals, such as speech signals. In this study the clinical course of a patient undergoing surgical removal of a vocal fold lesion has been analyzed using an adaptive mode decomposition approach to identify acoustic descriptors useful for noninvasively monitoring healing. The research started with the preliminary comparative evaluation of four algorithms: Intrinsic Time-scale Decomposition (ITD), Local Mean Decomposition (LMD), Robust Local Mean Decomposition (RLMD), and Iterative Filtering (IF). These methods were tested on both a numerical signal and an experimental benchmark to evaluate their robustness in the presence of noise and nonlinearity.
The best overall performances were of the IF algorithm , which was used to analyze four recordings of the sustained vowel /a/ recorded pre-surgery and during a year of recovery
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Informazioni aggiuntive
Corso di laurea
Classe di laurea
URI
![]() |
Modifica (riservato agli operatori) |
