Fine-grained differentiable physics: a yarn-level model for fabrics

Abstract

Differentiable physics modeling combines physics models with gradient-based learning to provide model explicability and data efficiency. It has been used to learn dynamics, solve inverse problems and facilitate design, and is at its inception of impact. Current successes have concentrated on general physics models such as rigid bodies, deformable sheets, etc, assuming relatively simple structures and forces. Their granularity is intrinsically coarse and therefore incapable of modelling complex physical phenomena. Fine-grained models are still to be developed to incorporate sophisticated material structures and force interactions with gradient-based learning. Following this motivation, we propose a new differentiable fabrics model for composite materials such as cloths, where we dive into the granularity of yarns and model individual yarn physics and yarn-to-yarn interactions. To this end, we propose several differentiable forces, whose counterparts in empirical physics are indifferentiable, to facilitate gradient-based learning. These forces, albeit applied to cloths, are ubiquitous in various physical systems. Through comprehensive evaluation and comparison, we demonstrate our model's explicability in learning meaningful physical parameters, versatility in incorporating complex physical structures and heterogeneous materials, data-efficiency in learning, and high-fidelity in capturing subtle dynamics. Code is available in: https://github.com/realcrane/Fine-grained-Differentiable-P hysics-A-Yarn-level-Model-for-Fabrics.git

Efficient structuring of the latent space for controllable data reconstruction and compression

“…Liang et al [2019] and Li et al [2022] presented approaches for cloth parameter estimation based on sheet-level differentiable cloth models. Gong et al [2022] introduced a differentiable physics model at a more fine-grained level, where yarns are modelled individually, thereby allowing to model cloth with mixed yarns and different woven patterns. Their model leverages differentiable forces on or between yarns, including contact, friction and shear.…”

Visual similarity simulation of slub denim based on image processing

“…17,18 Subsequently, some scholars established a finite element model of the fabric. 19 Then, the B-spline method [20][21][22] and the differentiable physical models [23][24][25] were proposed to simulate yarnlevel fabrics. Recently, the behaviors of the fabrics have been brought into focus.…”