Identifier of human emotions based on convolutional neural network for assistant robot

Received Sep 2, 2019 Revised Jan 4, 2020 Accepted Feb 23, 2020 Keywords: Autonomous robot Convolutional neuronal network Human emotions Service robot ABSTRACT This paper proposes a solution for the problem of continuous prediction in real-time of the emotional state of a human user from the identification of characteristics in facial expressions. The work presented in this paper aims to contribute to this area of research, particularly evaluating the performance of an automatic identification system of emotions in real-time on a laboratory robot assembled to perform tasks of assistance to human beings. [...]we conclude by discussing the lessons learned and how the system could be further developed. 2.PROBLEM FORMULATION The goal of this research is to develop a robust and high-performance software tool that allows the development of autonomous tasks of identifying human emotions by a small autonomous robot. There is also a face detector in dlib based on CNN (convolutional neural network), and although it has a better performance detecting faces in many angles (dlib.getjrontal_face_detector() only works well with frontal faces), we decided not to use it due to its high consumption of resources, which made it impossible to use in real-time on our assistant robot (the strategy based on CNN was a little more than 15 times slower in each frame than the strategy based on HOG).