 
Summary: · Learn priors on the spatiotemporal relationship between edges in the
alpha matte and the composite image
· Fit these priors to natural image statistics models to avoid overfitting
· These priors propagate alpha values along, but not across, edge surfaces in
both space and time
·
· We calculate the maximum a posteriori (MAP) estimate of the foreground
image F and the alphamatte given C and B:
·
· The MAP estimation is converted into an energy minimization by taking the
negative loglikelihood of the posterior, and noting that p(C) and p(B) do not
depend on the unknowns F and .
· This is calculated over 100,000's of variables using the wellhoned Matlab
optimizer fmincon. Derivatives of the energy function are computed using
Maple, and automatically written to C source code. An analytical sparse
Hessian is assembled from these derivatives allowing the use of efficient
sparse solvers.
Nicholas Apostoloff and Andrew W. Fitzgibbon
University of Oxford
{nema,awf}@robots.ox.ac.uk
