Deep Learning Models for Emotion Classification in Human Robot Interaction Platforms

Abstract

Human Robot Interaction (HRI) main purpose is to improve the communication between robots and people, in special the service robots which principal function is interacting with users. Service robots could be virtual or physical, such as a chatbot or humanoid robot. The increase of internet access and the use of online services have produced an exponentially use of chatbots. This situation generate people spending more time using this technology and trying to humanize it. Therefore, giving robots emotional capabilities have become an important issue in the field. For this reason, the purpose of this article is to analyzed and compared the performance of common deep learning techniques (CNN, RNN) that could be used as a emotion classifier for HRI platforms such a chatbots or humanoid robots. Two kind of input signals were evaluated: text and images of faces. In addition, different metrics were selected to evaluate the accuracy and time performance of the models.