← ArchivePaper2010
Stable Spaces for Real-time Clothing
Edilson de Aguiar, Leonid Sigal, Adrien Treuille, Jessica K. Hodgins
Data-driven conditional cloth model learned from simulation that enables real-time animation of thousands of garments with approximate collision resolution.
Abstract
We present a technique for learning clothing models that enables the simultaneous animation of thousands of detailed garments in real-time. This surprisingly simple conditional model learns and preserves the key dynamic properties of a cloth motion along with folding details. Our approach requires no a priori physical model, but rather treats training data as a "black box." We show that the models learned with our method are stable over large time-steps and can approximately resolve cloth-body collisions. We also show that within a class of methods, no simpler model covers the full range of cloth dynamics captured by ours. Our method bridges the current gap between skinning and physical simulation, combining benefits of speed from the former with dynamic effects from the latter. We demonstrate our approach on a variety of apparel worn by male and female human characters performing a varied set of motions typically used in video games ( e.g. , walking, running, jumping, etc. ).
How to read this
- Category
- Method: a data-driven (learned) clothing model
- Contributions
- A simple conditional model learned from simulation data that animates thousands of detailed garments in real time
- Stability over large time-steps with approximate cloth-body collision resolution and no a priori physical model
- Bridging skinning and physical simulation, combining skinning speed with simulation-like dynamic folding detail
- Context
- Positions itself between skinning and physical cloth simulation, treating simulation training data as a black box to learn a conditional dynamic model.
- Correctness
- Demonstrated on apparel for male and female human characters doing game-typical motions (walking, running, jumping); collisions are only approximately resolved, and the model is conditioned on its training distribution, so behavior outside that motion class is a fair concern.
- Clarity
- Accessible; the abstract frames the idea plainly, and a first pass conveys the skinning-vs-simulation bridge, with a second pass for the conditional-model details.
- How to read it
- Read for the conceptual placement between skinning and simulation and the stability claims; a second pass pays off if you need the learning setup and how collisions are approximated.
Builds on
Nothing in the archive, this is a starting point.
Related work
- A Pixel-Based Framework for Data-Driven Clothing 2020 / SCA
- Directing Cloth Draping through Blended UVs 2025 / SIGGRAPH
- Subspace Neural Physics: Fast Data-Driven Interactive Simulation 2019 / SCA
- PBNS: Physically Based Neural Simulation for Unsupervised Garment Pose Space Deformation 2021 / SIGGRAPH Asia
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →