Estimating Cloth Simulation Parameters From a Static Drape Using Neural Networks

Eunjung Ju,Myung Geol Choi

doi:10.1109/access.2020.3033765

Abstract

We present a neural network learning approach for estimating a set of cloth simulation parameters from a static drape of a given fabric. We use a variant of Cusick's drape, which is used in the fashion textile industry to classify fabric according to mechanical properties. In order to produce a large enough set of reliable training data, we first randomly sample simulation parameters using a Gaussian mixture model that is fitted with 400 sets of primary simulation parameters derived from real fabrics. Then, we simulate our modified Cusick's drape for each sample parameter set. To learn the training data, we propose a two-stream fully connected neural network model. We prove the suitability of our neural network model through comparisons of the learning errors and accuracy with other similar neural network and linear regression models. Additionally, to demonstrate the practicality of our method, we reproduce the drape shapes of real fabrics using the simulation parameters estimated from the trained neural networks.

Highlights

Cloth simulation is widely used in the fashion textile industry
Because the physical mechanism of the cloth simulation is different from the mechanism of real cloth, simulated drapes are often different from real drapes, even if you enter the precise mechanical properties
CLO3D was used as the simulator, and six simulation parameters that have a great influence on the formation of static drape were targeted

Summary

Introduction

Cloth simulation is widely used in the fashion textile industry. For example, the development costs of a new garment design can be dramatically reduced by predicting the final drapes through a simulation and compensating for problems. That means that the designer may need to measure the density, stretch stiffness, bending stiffness, and other properties of the fabric to determine the simulation parameters. KES [2] is the most representative instrument for measuring material It precisely measures the mechanical properties of fabric, such as tensile strength, bending stiffness, and shear stiffness, as well as simple physical properties, such as frictional force and density. Most of these properties are the same as the simulation parameters used in general cloth simulators. Our goal is not to measure the actual mechanical properties of a real fabric but to estimate the simulation parameters that reproduce the desired static drape. The downside is that it is difficult to recognize mechanical properties from a static image of a hanging drape [5], [8]

Objectives

Results

Conclusion