Generative and isoparametric geometric modeling of large-scale and multiscale microstructures

The paper argues that traditional geometric representations break down when microstructures need to span both fine detail and large physical domains. Explicit meshes are too memory-heavy, while implicit methods only stay compact when the shape can be described procedurally in a simple way. To address that, the authors introduce ExVCC, an extended volumetric Catmull-Clark spline that supports local refinement beyond tensor-product topology.

On top of ExVCC, they add shape-coding schemes and refinement rules that let geometry be generated on demand and evaluated hierarchically. They also propose an isoparametric representation where details across scales share the same parametric domain and spline basis, so edits can propagate automatically through the refinement hierarchy. For game developers, the interesting part is less about additive manufacturing itself and more about the modeling idea: compactly encoding dense detail while preserving relationships across LOD-like scales is...