| abstract |
Various systems have been presented to support the design of plant models. Most of them are using parameterized algorithms, rule-based descriptions or rewriting rules and grammars. Nature however often shows very irregular (branching-) structures that are very hard to describe with abstract mathematical models. Even for more regular plants it is still difficult to encode the structure in rules or a set of parameters. This method tries to overcome this by using a small set of unregistered images as direct input. Starting with a initial volume extracted from the input images we generate the geometry with a particle system. The system tries to mimic the feature of the natural branching system as an optimal transport system for nutriments and water. To achieve higher visual coherence with the input images the particle system is influenced by a graph system that was extracted from the input images as well. The project was funded by the Karl Steinbuch stipend of the 'Medien und Filmgesellschaft' and developed by Thomas Franken and Boris Neubert.
|
