A mobile application to report and detect 3D body emotional poses

Abstract

Most research into automatic emotion recognition is focused on facial expressions or physiological signals, while the exploitation of body postures has scarcely been explored, although they can be useful for emotion detection. This paper first explores a mechanism for self-reporting body postures with a novel easy-to-use mobile application called EmoPose. The app detects emotional states from self-reported poses, classifying them into the six basic emotions proposed by Ekman and a neutral state. The poses identified by Schindler et al. have been used as a reference and the nearest neighbor algorithm used for the classification of poses. Finally, the accuracy in detecting emotions has been assessed by means of poses reported by a sample of users.