Automated Facial Expression Analysis

Human Emotion Recognition Clips Utilised Expert System was designed to recognise and interpret facial expressions of the observed person in an automatic way [1]. Still, input to HERCULES has been manually supplied. Through integrating HERCULES into the Integrated System for Facial Expression Recognition (ISFER) a complete process of automatic analysis of facial expressions has been achieved [2]. ISFER forms a part of the project Automated System for Non-verbal Communication [3] which is an ongoing project at the Knowledge Based Systems department of Delft University of Technology. The theoretical formulation of the implemented facial expression recognition has been acquired from FACS [4]. The interpretation of the recognised facial expression is currently based [1][7] on the recognition of so-called six basic emotions defined by Ekman [5][6]: happiness, sadness, fear, surprise, disgust and anger. Validation of the implemented knowledge and performing of the facial expression analysis in a completely automated way, are the main topics of this paper.