Introduction

Here we give a brief introduction to statistical appearance models. It describes how one can build a model of the appearance of an object given a set of manually labelled images of this object. Our eventual aim is to extend such a model to be able to model objects in video sequences.

The second part of the chapter describes how an active appearance model is used to track the face through a long video sequence, in the face tracking module. This is a description of the first module shown in the overview of the system (see figure 3.1).

Figure 3.1: Overview of the components of the model. This chapter explains how the face tracking module works.