Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Emotion Recognition System for Human-Robot Interface: Comparison of Two Approaches
Ist Teil von
Interactive Collaborative Robotics, 2017, p.20-28
Ort / Verlag
Cham: Springer International Publishing
Erscheinungsjahr
2017
Quelle
Alma/SFX Local Collection
Beschreibungen/Notizen
This paper describes a system for automatic emotion recognition developed to enhance the communication capabilities of an anthropomorphic robot. Two versions of the classification algorithm are proposed and compared. The first version is based on a classic approach requiring the action unit estimation as a preliminary step to emotion recognition. The second version takes advantage of convolutional neural networks as a classifier. The designed system is capable of working in real time. The algorithms were implemented on C++ and tested on an extensive face expression database as well as in real conditions.