Estimation of parameters in a planar segment process model
We consider a model for planar segment process with interactions arising from intersections among segments. The process is given by a density w.r.t. a Poisson process and a reference orientation distribution is involved. The problem of parameter estimation is solved first in a fully parametric case and secondly in a semiparametric way where the orientation distribution is not parametrized. The methods are tested in simulated data. Finally a potential application to the images of stress fibres in stem cells is discussed.
Joint work with Jakub Večeřa.