← ArchivePaper2021

GarMatNet: A Learning-Based Method for Predicting 3D Garment Mesh with Parameterized Materials

Two-stream network predicts body-fitted garment mesh deformation conditioned on pose and parameterized fabric material properties.

Open the paper ↗Trace in the lineage graph

Abstract

Recent progress in learning-based methods of garment mesh generation is resulting in increased efficiency and maintenance of reality during the generation process. However, none of the previous works so far have focused on variations in material types based on a parameterized material parameter under static poses. In this work, we propose a learning-based method, GarMatNet, for predicting garment deformation based on the functions of human poses and garment materials while maintaining detailed garment wrinkles. GarMatNet consists of two components: a generally-fitting network for predicting smoothed garment mesh and a locally-detailed network for adding detailed wrinkles based on smoothed garment mesh. We hypothesize that material properties play an essential role in the deformation of garments. Since the influences of material type are relatively smaller than pose or body shape, we employ linear interpolation among different factors to control deformation. More specifically, we apply a parameterized material space based on the mass-spring model to express the difference between materials and construct a suitable network structure with weight adjustment between material properties and poses.

How to read this

Category: Method: learning-based garment deformation conditioned on material
Contributions: GarMatNet, a two-stream network predicting garment deformation from human pose and parameterized garment material
A generally-fitting network for smoothed garment mesh plus a locally-detailed network that adds wrinkles
A parameterized material space based on a mass-spring model, with linear interpolation to control material-driven deformation
Context: Sits in the learning-based garment animation line of work, building on virtual-try-on clothing animation such as Santesteban et al.Builds on: Learning-Based Animation of Clothing for Virtual Try-On
Correctness: Operates under static poses and assumes material influence is smaller than pose or body shape, justifying linear interpolation across materials; this interpolation assumption and the static-pose setting are the limitations to keep in mind.
Clarity: Accessible (MIG); a first pass conveys the two-stream idea, a second pass clarifies the material parameterization.
How to read it: Read for how material is parameterized and injected, and the coarse-to-detail two-network split; second pass if you care about the mass-spring material space.

Builds on

Learning-Based Animation of Clothing for Virtual Try-On 2019

Built upon by

Nothing yet.

Related work

N-Cloth: Predicting 3D Cloth Deformation with Mesh-Based Networks 2022 / CGF
Motion Guided Deep Dynamic 3D Garments 2022 / SIGGRAPH Asia
Learning-Based Animation of Clothing for Virtual Try-On 2019 / CGF
PBNS: Physically Based Neural Simulation for Unsupervised Garment Pose Space Deformation 2021 / SIGGRAPH Asia

Keywords

garmatnet garment deformation prediction parameterized materials mass-spring model generally-fitting network locally-detailed network multilayer perceptron laplacian smoothing smpl tailornet physically-based simulation per-vertex displacement wrinkles weight controller marvelous designer mean square error

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 →