Facial Expression Analysis

There has recently been high interest in affective computing, especially in interfaces which can analyse their users’ emotional state. Automatic emotion recognition in faces is a hard problem, requiring a number of pre-processing steps which attempt to detect or track the face, to locate characteristic facial regions such as eyes, mouth and nose on it, to extract and follow the movement of facial features, e.g., characteristic points in these regions, or model facial gestures using anatomic information about the face. In this work we present a methodology for analysing both primary and intermediate expressions. This is performed through a system which, after a skin color segmentation and the extraction of the face and of the facial points, translates FP movements into FAPs and reasons on the latter to recognize the underlying emotion in facial video sequences. The developments described in the current work are being extended and validated in the framework of the IST ERMIS project.