Análise sensorial: desenvolvimento de um sistema computacional para verificação de aceitação de produtos

Abstract

Sensory analysis assesses consumer satisfaction with products by perceiving emotional reaction through feeling, in particular, facial expressions. However, it often uses manual and expensive processes, with human errors, from the wrong understanding of reaction to registration. This article presents a computational system that assists the analysis of photosensitive products. The system uses Arduino and software to control LED-RGB lights, manages experimental sessions, framework for recognizing facial expressions by Convolutional Neural Networks, multiplatform desktop graphical interface and database. As a result, there is a solution that guarantees the security and reliability of information.