Reconocimiento multimodal de emociones humanas en un ambiente inteligente

Abstract

Affective computing allows computers to recognize and generate emotions to improve human-machine interaction. In this field of research, this project presents a multimodal emotion recognition model in an ambient intelligence, oriented to a smart classroom. Initially, it was proposed an architecture for the implementation of the management of emotions in an ambient intelligence (AmI) which consists of physical, agents and services layer. The emotion recognition model designed is multimodal since the determination of emotions is done through an analysis of facial expressions and a text analysis. On the other hand, to analyze the model has taken as a reference the six basic emotions: happiness, sadness, anger, fear, disgust and surprise proposed by Paul Ekman. It was done a study of the tools that help the multimodal emotion recognition and then were selected the best ones. The multimodal model has been tested in two cases of study applied in a classroom, while students visualize a video and then an essay of the visualized video.