|
PresentationsFast marching to branching morphologiesBiology faculty, Lomonosov Moscow State University 1Department of molecular neuroscience, IBCh RAS 2Department of molecular neuroscience, IBCh RAS Branching self-similar structures providing for transport of nutrients, air, or information are widespread in nature. In computational biology and miomimetics applications, a need to model such structures often arises. We propose a framework for simulating branching morphologies, in which branches are represented by merging gradient descent paths on travel time maps generated by the fast marching method over stochastic speed fields. Feedback between the current tree and the speed profile and the wave source allows generating a wide range of realistic morphologies. Using multiple wave sources in the fast marching method allows generating tiling networks with competing non-overlapping territories, reminiscent of brain astrocyte networks.
|
