next up previous index
Next: Encoding transition probabilities Up: Variable length Markov model Previous: Variable length Markov model   Index





Introduction

Figure 6.1: Overview of the components of the model. This chapter explains how to model temporal relationship between pathlet models.
\includegraphics[width=145mm,keepaspectratio]{frameworkoverviewvlmm.eps}

The previous chapter demonstrated how to extract a sequence of pathlet models from a trajectory of active appearance model parameters. We now describe how a variable length Markov model can be used to model the sequence.

In the following, we describe the VLMM along with the different components that can be used within its learning scheme. Figure 6.1 shows how the VLMM module fits into the whole framework of our model.



franck 2006-10-01