In this chapter, we have shown how we can use the active appearance model to track a face in a video sequence.
We have applied that tracking to 3 different video sequences. All the video sequences have been tracked entirely to generate sequences of appearance parameter vectors.
The following chapters will describe how we train models to generate trajectories through parameter space similar to these sequences. This allows us to synthesise convincing faces that mimic the behaviour of the face in the original video sequence.