SDF Grid Mean Curvature Node

The SDF Grid Mean Curvature node applies mean curvature flow smoothing to a Signed Distance Field (SDF) grid. This operation evolves the surface of the field over time based on its mean curvature, causing high-curvature (sharp or noisy) regions to smooth out more rapidly than flatter areas.

Unlike simple averaging or Laplacian smoothing, mean curvature flow adapts to the geometric features of the surface, providing a more natural, shape-preserving smoothing process. It can be used to refine complex SDF surfaces, remove small artifacts, and create organically blended transitions between features.

This node is particularly effective after boolean operations or mesh-to-SDF conversion, where the resulting fields may contain high-frequency noise.

Inputs

Grid

The input signed distance field to smooth. Typically, the grid represents a surface or volume generated from a mesh or procedural operation.

Iterations

The number of mean curvature flow iterations to apply. Each iteration progressively smooths the surface, with higher values resulting in a more pronounced smoothing effect. A small number of iterations will gently polish the surface, while larger values can significantly alter its shape.

Outputs

Grid

The resulting SDF grid after mean curvature flow smoothing. The output surface has reduced high-curvature noise and smoother transitions, while retaining major structural features of the original shape.