Two concomitant issues are addressed: facial expressions from 3D images and face and head gestures from 2D videos. Human face is an effective carrier of various types of communicative and emotional information, whose understanding and recognition is essential to enhance any man-machine communication system. 3D can capture true facial surface data and is less disturbed by illumination and head pose changes. The best performance is achieved by a fusion of 2D and 3D modalities. Recognition of facial expressions in videos can be based on face landmark dynamics and face appearance changes. Sequence processing carried out by Hidden Markov Models or by Hidden Conditional Random Fields achieves very satisfactory recognition performances when tested in various test corpora. More...

Augmented Reality is on the top of technology hype curve, and fancy demos appear in a growing speed. This talk will look beyond the hype and discuss what it is good for in reality, and what are the trends and bottlenecks of future development.