Selective and Dynamic Cloth Fold Smoothing with Collision Resolution

How to read this

Category

Production technique: post-simulation cloth fold smoothing and collision cleanup

Contributions

• Per-vertex fold detection via a position-based method (original vs internally smoothed mesh) or an angle-based method (angular change across shared edges), producing spread and spatially smoothed fold weights
• Selective Laplacian, Taubin, or Delta Mush smoothing driven by those weights to blend between the original and a fully smoothed mesh, giving artists emphasis and de-emphasis control
• A collision resolution pass that pushes interpenetrating vertices out of the body and smooths offsets, implemented as a multi-threaded custom node after the cloth sim step

Context

Builds on the simulation-of-folds-and-wrinkles tradition (Bridson et al., Simulation of Clothing with Folds and Wrinkles), operating as an artist-controllable post-process rather than a new solver.Builds on: Simulation of Clothing with Folds and Wrinkles

Correctness

Presented as a DigiPro production technique demonstrated qualitatively in a pipeline node; assumes a usable simulated mesh exists and that selective smoothing will not erase desired detail, so results are tuned by artists rather than benchmarked.

Clarity

Practical and accessible; a first pass conveys the detect-weight-smooth-resolve workflow, with details in the smoothing-operator and collision sections.

How to read it

First pass for the pipeline-stage role and the fold-detection-to-weight idea; revisit the smoothing operators (Laplacian/Taubin/Delta Mush) and the collision pass if integrating into a cloth pipeline.