Keywords
ITL (KT0), reconocimiento; técnica de clasificación; emoción; voz, recognition; classification technique; emotion; voice
Reconocimiento de emociones a través de la voz
Abstract
En el presente proyecto se ha realizado un estudio comparativo de unas técnicas de clasificación de datos utilizando un conjunto de características de la señal de la voz con el fin de encontrar las propiedades de la voz más significativas a la hora de reconocer emociones y, a su vez, definir la técnica con la que se obtienen mejores resultados. Para ello, se han implementado tres soluciones: un open-software basado en máquinas de soporte vectorial (SVM), unos mapas autoorganizados (SOM) y los SOMs complementados con árboles de decisión.
In this project, a comparative study of data classification techniques using a set of characteristics of the voice signal has been carried out. The main objective has been to find the most significant properties of the voice when recognizing emotions and identify the technique with which best results are obtained. Therefore, three solutions have been implemented: an open-software based on vector support machines (SVM), self-organized maps (SOM) and SOMs complemented with decision trees.
In this project, a comparative study of data classification techniques using a set of characteristics of the voice signal has been carried out. The main objective has been to find the most significant properties of the voice when recognizing emotions and identify the technique with which best results are obtained. Therefore, three solutions have been implemented: an open-software based on vector support machines (SVM), self-organized maps (SOM) and SOMs complemented with decision trees.
Información detallada
Materias, derechos, colecciones e identificadores