Estimating Cloth Simulation Parameters From a Static Drape Using Neural Networks

Ju, Eunjung; Choi, Myung Geol

doi:10.1109/access.2020.3033765

Cited by 23 publications

(27 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Friction coefficients have been estimated using reflectance values [ZDN16] or dynamic videos of cloth sliding through a surface [RRBD * 20] Instead of regressing the parameters, Huber et al [HEW17] find the most similar cloth in a database using motion descriptors. A different approach only using a single image of the Cusick drape was followed by Ju et al [JC20], but it requires a 360 • scan to reconstruct the target cloth, and a manually fitted Bezier curve to obtain the feature vector. In contrast, we just require a depth map that can be captured easily.…”

Section: Parameter Estimation Methodsmentioning

confidence: 99%

“…Accurate fabric parameter acquisition systems require specialized and expensive devices [Kaw80,Min95,CPGE90], which are often slow and need skilled operators. Existing casual capture setups use input video sequences [BTH * 03, BXBF13,YLL17] or, even if they take a single image, might require manual user input [JC20]. In this paper, we present a casual capture system that only requires taking two depth images of the textile posed in a static drape.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

How Will It Drape Like? Capturing Fabric Mechanics from Depth Images

Rodríguez-Pardo

Prieto-Martin

Casas

et al. 2023

Computer Graphics Forum

View full text Add to dashboard Cite

We propose a method to estimate the mechanical parameters of fabrics using a casual capture setup with a depth camera. Our approach enables to create mechanically-correct digital representations of real-world textile materials, which is a fundamental step for many interactive design and engineering applications. As opposed to existing capture methods, which typically require expensive setups, video sequences, or manual intervention, our solution can capture at scale, is agnostic to the optical appearance of the textile, and facilitates fabric arrangement by non-expert operators. To this end, we propose a sim-to-real strategy to train a learning-based framework that can take as input one or multiple images and outputs a full set of mechanical parameters. Thanks to carefully designed data augmentation and transfer learning protocols, our solution generalizes to real images despite being trained only on synthetic data, hence successfully closing the sim-to-real loop. Key in our work is to demonstrate that evaluating the regression accuracy based on the similarity at parameter space leads to an inaccurate distances that do not match the human perception. To overcome this, we propose a novel metric for fabric drape similarity that operates on the image domain instead on the parameter space, allowing us to evaluate our estimation within the context of a similarity rank. We show that out metric correlates with human judgments about the perception of drape similarity, and that our model predictions produce perceptually accurate results compared to the ground truth parameters.

show abstract

Section: Parameter Estimation Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

How Will It Drape Like? Capturing Fabric Mechanics from Depth Images

Rodríguez-Pardo

Prieto-Martin

Casas

et al. 2023

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…The sample thickness was determined using a thickness tester (No. 20465, Mitutoyo Co., Kawasaki, Japan) according to the ASTM D 5729-9 standard . The bending distance and length were measured using a bending test kit, and the bending stiffness was derived using eq according to the KS K ISO 9073-7 standard …”

Section: Methodsmentioning

confidence: 99%

“…20465, Mitutoyo Co., Kawasaki, Japan) according to the ASTM D 5729-9 standard. 29 The bending distance and length were measured using a bending test kit, and the bending stiffness was derived using eq 1 according to the KS K ISO 9073-7 standard. 30 Table 1 summarizes the KES-FB test conditions.…”

Section: Methodsmentioning

confidence: 99%

Comparison of the Silhouette of Virtual Clothes by Fabric Characteristics of Nylon Fabric for the Utilization of Virtual Clothes

Shim

Koo

et al. 2022

ACS Omega

View full text Add to dashboard Cite

In this study, various nylon fabrics were investigated to measure the physical properties of the actual fabric using Kawabata systems and CLO fabric kits, and the correlation between the physical and drape properties of the fabric was statistically analyzed. Subsequently, the CLO fabric kit results were used as basic data for the CLO three-dimensional (3D) virtual clothing program, and the drapeability of the virtual fabrics was measured. The factors with the most significant effect on the drape properties were analyzed by comparing the drape measurement of the real fabrics through the variable measurement of various fabrics and correlation analysis with the drape properties, and attempts were made to reduce the difference between real fabrics and their virtual implementations. The correlation analysis results revealed that both the Kawabata system and CLO fabric kit results indicated that the bending and shear properties exhibited the highest correlation with the drape ratio. In addition, the comparison of the real and virtual fabric images revealed that they exhibited a similar morphology. Furthermore, the results confirmed that the difference between the real and virtual fabrics could be reduced as much as possible when the values measured using the Kawabata system and CLO fabric kit were converted into a database. Subsequently, the differences between virtual fabrics with various thicknesses were confirmed through CLO 3D simulation. The findings of this study are expected to be used as basic data for building fabric database.

show abstract

“…Ju et al [19] offer options, using different methods of draping the fabric and threedimensional scanning. Such draping methods are hanging drape, Cusick's drapemeter, and modified Cusick's drapemeter.…”

Section: Introductionmentioning

confidence: 99%

Drape of Composite Structures Made of Textile and 3D Printed Geometries

et al. 2022

View full text Add to dashboard Cite

Applications of 3D printing in the fashion industry have continued to attract interest from academia and industry in order to improve and add functionalities to products. Among these applications, an interesting one is 3D printing on textile fabric. Composite structures created by 3D printing and textile fabric change a drape by improving or worsening its appearance. The scope of this work is to evaluate the effect of various 3D printed geometries on textile fabric regarding fabric drapes. The drape coefficient of the created composite structure is evaluated using a drape tester built according to EN ISO 9073-9. The results taken are compared with an algorithm developed for determining drape parameters and 3D form representation using color digital images and their image histograms. The measured values of the drape coefficient are close, with a minimal difference, up to 4%. The 3D printed patterns show a significant effect on the drape coefficient of textile fabrics by depicting another way to modify fabric drapes and create complex shapes by using less material. This can be seen as an advantage in the fashion industry where complex geometries can be added to textile fabrics, while changing fabric drape and product personalization and adding functionalities for garments and technical textiles.

show abstract

Estimating Cloth Simulation Parameters From a Static Drape Using Neural Networks

Cited by 23 publications

References 24 publications

How Will It Drape Like? Capturing Fabric Mechanics from Depth Images

How Will It Drape Like? Capturing Fabric Mechanics from Depth Images

Comparison of the Silhouette of Virtual Clothes by Fabric Characteristics of Nylon Fabric for the Utilization of Virtual Clothes

Drape of Composite Structures Made of Textile and 3D Printed Geometries

Contact Info

Product

Resources

About