The human face is a very powerful medium of communication, which can express a wide range of meanings. These different meanings can obviously be expressed at different time but also concurrently which produce complex facial displays. One of my interest is to reproduce these kind of complex facial displays with synthetic faces.

My first try to produce these types of expressions was by developing an emotion representation, the DER, that is composed of two types of emotions: Secondary Emotions and Mood. Each one of these emotion types is generating its own facial expression and both facial expressions are merged together to produce a final facial display. Complex facial displays could be produced by contradictory emotional states.

The two facial displays below are produced by the same intensity of Secondary Emotion but with different mood states. The mood state for the picture on the left is neutral where the mood state for the picture of the right is "bad".

The differences between these two pictures are quite subtle. More work need to be carried out to make the differences more visible and to evaluate their recognition and influence.

Just by curiosity, there is one question I like to ask:
If you have to buy a car from one of these two characters, which one will you choose to make the deal with? If you have any opinion, please, email me at e.tanguy@cs.bath.ac.uk.