ABSTRACT
Facial expressions are among the most important non-verbal communication cues for human-to-human interaction. They are one of the few windows available to an observer on the emotions felt by an individual. Therefore, in the context of better understanding human-to-human communication and to enable natural human-computer interactions, automatic facial expression recognition has been deeply investigated in IM2. In this chapter we will describe the main processing steps involved in feature extraction for facial expression recognition, as well as the machine learning tools developed for two specific goals: recognition of the facial expressions on static images, and modeling of the perception of facial expressions by a large cohort of observers, from different geographical and cultural origin.
