Sistema de reconocimiento de emociones en la voz para un tutor inteligente.

Abstract

The automatic recognition of human emotion in
speech aims to recognize the emotional state of a
speaker from the voice signal. The voice signal
contains information about age, sex, social status,
accent and emotional state of a speaker, as well as
linguistic information.
The present work describes the advances obtained
to recognize affective states based on specific
characteristics of the speech and will implement a
system that detects the emotion of anger. The
database of Berlin is used, which is free and many
researchers have carried out their work and where
the results reported have not exceeded 82%
The process of initiation when the sound of a word
is captured by the computer that makes a recording
or the reading in WAV format, the act followed the
signal is processed and normalized, after obtaining
the same Fast Fourier Transform (FFT), finally, by
means of the error coefficients, the result is
compared in the database that contains the patterns
of a recognize. It addresses the use of the correlation
technique as a method of recognition of emotions in
the voice (REV).