Thesis ESTIMACIÓN DEL FLUJO AÉREO GLOTAL A PARTIR DE UNA SEÑAL DE ACELERÓMETRO UTILIZANDO REDES NEURONALES
Date
2017
Authors
FARÍAS CORREA, ROBERTO EUGENIO
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Muchos trastornos de la voz comunes son condiciones crónicas o recurrentes que puedenser el resultado del uso inapropiado o excesivo de la voz (patrones defectuosos y/o abusivosdel comportamiento vocal). A este tipo de trastornos de la voz se les refiere comúnmentecomo hiperfunción vocal, los cuales pueden llevar amplias consecuencias ya sea en losocial, profesional o personal.Con el fin de proporcionar una solución no invasiva para la prevención y el monitoreode este tipo de trastornos, han estado en desarrollo plataformas como el monitor de lasalud de la voz, el cual busca a través del análisis de la aceleración proporcionada por unaacelerómetro posicionado en la base del cuello, determinar medidas de relevancia médica(como lo es el flujo aéreo glotal) para el diagnóstico y prevención de trastornos de la voztales como la hiperfunción vocal.Este proyecto busca estudiar e implementar un algoritmo basado en un modelo de redesneuronales, el cual permita la estimación del flujo aéreo glotal utilizando una señal provenientede un acelerómetro adherido a la piel del cuello de la persona. El objetivo es poner aprueba un acercamiento que use un modelo de redes neuronales y cuantificar su desempeñoen comparación con otros acercamientos al problema, y de esta manera ayudar buscar unaposible mejora en la eficacia de dispositivos ambulatorios que lo requiran, como el monitorde la salud de la voz.
Many common voice disorders are chronic or recurrent conditions that may be the result ofimproper or excessive use of voice (defective and/or abusive patterns of vocal behavior).These types of speech disorders are commonly referred to as vocal hyperfunction, whichcan lead to farreaching social, professional or personal consequences.In order to provide a non-invasive solution for the prevention and monitoring of this type ofdisorders, platforms such as the voice health monitor have been developed, which it looksfor through the analysis of the acceleration provided by an accelerometer positioned at thebase of the neck, determine measures of medical relevance (such as glottal airflow) for thediagnosis and prevention of voice disorders such as vocal hyperfunction.This project seeks to study and implement an algorithm based on a neural network model,which allows estimation of glottal airflow using a signal from an accelerometer attached tothe skin of the person’s neck. The objective is to test an approach using a neural networkmodel and quantify its performance in comparison with other approaches to the problem,and thus help to look for a possible improvement in the effectiveness of ambulatory devicesthat require it, such as the Voice Health Monitor.
Many common voice disorders are chronic or recurrent conditions that may be the result ofimproper or excessive use of voice (defective and/or abusive patterns of vocal behavior).These types of speech disorders are commonly referred to as vocal hyperfunction, whichcan lead to farreaching social, professional or personal consequences.In order to provide a non-invasive solution for the prevention and monitoring of this type ofdisorders, platforms such as the voice health monitor have been developed, which it looksfor through the analysis of the acceleration provided by an accelerometer positioned at thebase of the neck, determine measures of medical relevance (such as glottal airflow) for thediagnosis and prevention of voice disorders such as vocal hyperfunction.This project seeks to study and implement an algorithm based on a neural network model,which allows estimation of glottal airflow using a signal from an accelerometer attached tothe skin of the person’s neck. The objective is to test an approach using a neural networkmodel and quantify its performance in comparison with other approaches to the problem,and thus help to look for a possible improvement in the effectiveness of ambulatory devicesthat require it, such as the Voice Health Monitor.
Description
Catalogado desde la version PDF de la tesis.
Keywords
FLUJO AEREO GLOTAL , HIPERFUNCION VOCAL , MODELO DE REDES NEURONALES , TRASTORNOS DE LA VOZ