The first step is to update the spatial model associated with each
blob using the blob's dynamic model, to yield the blob's predicted
spatial distribution for the current image:
where the estimated state vector 
includes the blob's
position and velocity, the observations Y are the mean spatial
coordinates of the blob in the current image, and the filter 
is
constructed assuming simple Newtonian dynamics. Smaller blobs near
the person's extremities (e.g., head, hands, and feet) are assumed to
have less inertia than the larger blobs that describe the person's body.