Catálogo de publicaciones - tesis

Compartir en
redes sociales

Título de Acceso Abierto

Modelado estocástico de la fonación y señales biomédicas relacionadas: Métodos en espacio de estados aplicados al análisis estructural, al modelado de la fonación y al filtrado inverso

Gabriel Alejandro Alzamendi María Eugenia Torres Virginia Ballarín Leandro Di Persia Agustín Gravano Humberto Torres Gastón Schlotthauer

acceptedVersion.

Resumen/Descripción – provisto por el repositorio digital
The voice and related biomedical signals carry information characterizing the phonation. The accurate extraction of this information becomes beneficial in medicine and other fields. In this Thesis, new methods for modeling the voice production and related biomedical signals, combining the stochastic modeling and digital signal processing, are proposed and evaluated. First, period and amplitude series extracted from sustained vowels are investigated applying the fundamental concepts of voice perturbations and fluctuations, and then a new sustained vowel synthesis method is proposed considering perturbations controlled with two acoustical parameters relevant in voice therapy. Next, state space structural models for studying period and amplitude series are developed. The aim of these models is to represent a signal assuming that it is composed of simple elements with a straightforward interpretation, under the hypothesis that these phenomena behave as non-stationary stochastic processes. Combining the structural models and state space methods, the structural analysis method is successfully implemented. Finally, new state space based methods for phonation modeling and voice inverse filtering are proposed. For that, a stochastic non-stationary difference equation of glottal function is developed, taking into account the perturbations and aperiodicities in this signal. According to this, a phonation model suitable for accurate and flexible vocal sounds representation is formulated. Using this model, a state space based inverse filtering method is proposed allowing the join optimal estimation of the glottal function and the vocal tract filter.
Palabras clave – provistas por el repositorio digital

Modeling of phonation and related biomedical signals; Analysis and synthesis of voice signals; State space methods; Modeling of non-stationary stochastic phenomena; Structural analysis of biomedical signals; Voice inverse filtering; Modelado de la fonación y señales biomédicas; Análisis y síntesis de señales de voz; Métodos en espacio de estados; Modelado de fenómenos estocásticos no estacionarios; Análisis estructural de señales biomédicas; Filtrado inverso de la voz

Disponibilidad
Institución detectada Año de publicación Navegá Descargá Solicitá
No requiere 2016 Biblioteca Virtual de la Universidad Nacional del Litoral (SNRD) acceso abierto

Información

Tipo de recurso:

tesis

Idiomas de la publicación

  • español castellano

País de edición

Argentina

Fecha de publicación