Análise facial de emoções utilizando redes neurais no contexto de uma sala de aula inteligente

Abstract

Smart classrooms emerge as an innovation in the field of education, aiming to improve teaching and assisting both teachers and students in classrooms. By equipping educational spaces with sensors and artificial intelligence it is possible to analyze an environment and promote feedback on the student?s motivation and interest in the class. The information produced can then be used by the teacher to understand about the progress of his or her classes, and thus, better structure them. Within this context, this work presents the concepts of intelligent classroom, computer vision and neural networks, necessary for the construction of software capable of analyzing the images of students in the classroom in order to identify patterns of emotions. In order to carry out the development of a software, the processes of image acquisition and processing, definition and implementation of a neural network were studied, culminating in the performance of tests. In this context, this work presents the proposal of applying a convolutional neural network to analyze images of students facial expressions in a teaching environment and recognize the emotions that are being expressed. Tests carried out confirmed the feasibility of using convolutional neural networks to recognize emotions in facial expressions, as related studies previously pointed out. It is understood that, in this way, it will be possible to integrate the proposed software to an IoT (internet of things) system that will constitute a relevant component to a classroom, thus adding a preliminary level of intelligence to it (sic).