Fast Voxelization and Level of Detail for Microgeometry Rendering

Abstract

Many materials show anisotropic light scattering patterns due to the shape and local alignment of their underlying micro structures: surfaces with small elements such as fibers, or the ridges of a brushed metal, are very sparse and require a high spatial resolution to be properly represented as a volume. The acquisition of voxel data from such objects is a time and memory-intensive task, and most rendering approaches require an additional Level-of-Detail (LoD) data structure to aggregate the visual appearance, as observed from multiple distances, in order to reduce the number of samples computed per pixel (E.g.: MIP mapping). In this work we introduce first, an efficient parallel voxelization method designed to facilitate fast data aggregation at multiple resolution levels, and second, a novel representation based on hierarchical SGGX clustering that provides better accuracy than baseline methods. We validate our approach with a CUDA-based implementation of the voxelizer, tested both on triangle meshes and volumetric fabrics modeled with explicit fibers. Finally, we show the results generated with a path tracer based on the proposed LoD rendering model.

Citation

@article{Fabre:2026:FastVoxelization, 
  author    = {Fabre, Javier and Castillo, Carlos and Rodriguez-Pardo, Carlos and Lopez-Moreno, Jorge},
  title     = {Fast Voxelization and Level of Detail for Microgeometry Rendering},
  journal   = {The Visual Computer},
  year      = {2026},
  publisher = {Springer}
}

Description

We present a GPU-based method to efficiently render complex microgeometry using a hierarchical voxel representation. Each voxel encodes a directional microflake distribution (SGGX), allowing accurate modeling of anisotropic materials beyond simple density-based approaches.

Our method performs fast parallel voxelization and builds a multi-resolution hierarchy by aggregating statistical distributions instead of averaging values, preserving appearance across scales. The hierarchical process merges closests SGGX functions to reduce the resulting error (Jensen-Shannon divergence from GT):

Our representation preserves directional detail that is lost in traditional voxel methods. As shown below, SGGX-based voxels preserve anisotropy, producing more accurate results than traditional approaches.

Render comparison against Ground Truth data

Different levels of detail can be used depending on distance, reducing cost while keeping the material's appearance consistent.