Quantitative analysis of non-verbal communication competence

Abstract

Oral communication is an escencial part of human life. Efeective communication skills are required not only professionally, but also in our daily life. Due to its importance in our daily activities it is essential to find methods that improve our communication skills. Several studies in the field of psychology on verbal and non-verbal communication has been done in the past decades, however from the point of view of Artificial Intelligence we are still at the begining of a long way to go. In this project, we propose a method that is able to extract behavioral cues from subject's presentations based on the analysis of multi modal data provided by the Kinect camera and applying different statistical classifiers that will allow the system to predict the quality of the presentation. All the experiments were took from students at Universitat de Barcelona while defending its class works.