The interaction between humans and computers lacks intuition; mostly it is restricted to traditional input and output devices. Therefore, integrating interpersonal interaction mechanisms will abolish this shortcoming. The proposed algorithms robustly localize facial features, seamlessly track them through image sequences, and determine features for facial expression interpretation.
The general research statement of this thesis is that model-based techniques have great potential to interpret images. Unfortunately, remaining challenges, such as the initial model parameterization, still forms major obstacles for a comprehensive application in real-world scenarios. In order to tackle this, we consider the objective function as the most important component involved. We propose a methodology that learns highly robust objective functions from annotated example images. This procedure lays the foundation for a general application of model-based image interpretation, because it is easy to use and does not require computer vision expertise.
«
The interaction between humans and computers lacks intuition; mostly it is restricted to traditional input and output devices. Therefore, integrating interpersonal interaction mechanisms will abolish this shortcoming. The proposed algorithms robustly localize facial features, seamlessly track them through image sequences, and determine features for facial expression interpretation.
The general research statement of this thesis is that model-based techniques have great potential to interpret...
»