Andrei Sherstyuk
Department of Computer Science
Monash University
There are few objects in the natural world that can be adequately represented by linear, circular or planar primitives. We show that in the domain of implicit modeling such primitives are very useful. Lines, arcs and triangles, when used as skeletons of convolution surfaces, produce well-structured and compact datasets that are easy to manipulate and animate.
We discuss a number of new modeling techniques and a number of datasets, that yield visually convincing shapes, resembling various forms of marine life. We also present volumetric textures that alter the actual geometry of implicit surfaces and make them more realistic.
Keywords: natural forms, geometric modeling, convolution surfaces, implicit surfaces.