09.09.2022

Digital yarn for real socks - ISTA scientists from the field of visual computing transfer yarn-based fabric simulation

The fact that Pixar and Disney characters don't run around naked is part of the legacy of the Wojtan group at ISTA. Their visual calculations of knitted textiles capture the complex behavior of garments. Now they are transferring their knowledge to a real-world application, benefiting the textile industry in the development of new fabrics. They are presenting their results at the prestigious SIGGRAPH conference, thus contributing to ensuring that people won't be running around naked in the future.

Narain et al. (2021)

Computer simulation of fabrics at the level of individual threads. Computer scientists on Chris Wojtan's team modeled the dynamics of knitted fabrics by applying the physical laws governing yarns. Their latest project aims to help the textile industry develop new fabrics.

Compare today's animated films with those from twenty years ago. Undoubtedly, you'll recognize the groundbreaking advances made in the field of visual data processing. This is the hard work of computer scientists like doctoral candidate Georg Sperl and his supervisor, Professor Chris Wojtan, at the Institute of Science and Technology Austria (ISTA). But it's not just animation artists who travel to the annual SIGGRAPH conference to learn about the latest findings. Technology and industry leaders are also there, searching for powerful algorithms. "For this project, we applied our efficient and accurate visualization programs to real-world problems and based the investigation on real data from the textile industry," Sperl summarizes the collaboration with the Spanish company SEDDI and the US company Under Armour.

From basic research to technological application

What's special about Sperl's simulations of knitted fabrics is the yarn-based approach. Instead of using a grid that only represents the overall properties of the material, it considers each individual yarn and its physical characteristics. This offers better control and captures more of the complexity of a moving knitted sweater. Yet the clever algorithm remains efficient. Until now, no one had applied yarn-based simulation to real-world industrial data. "We were curious to see if it worked. Real data is tricky. Many parameters are unknown. But the results show that it's possible and offers many advantages," says Sperl, giving an example:

Imagine a textile company wants to add a new fabric to its portfolio but doesn't know its properties—how the fabric twists, moves, and stretches. Note that a knitting pattern significantly and in complex ways alters the fabric's behavior. Now, the company can provide data on various knitting patterns made from the same yarn. With the new method, they can then calculate a yarn model that captures not only the dynamics of the provided pattern but also numerous other patterns using that yarn. Instead of producing and testing every possible option, the company can simulate the properties in advance. Such virtual testing saves resources.

“The textile industry is enormous, and simulation-based approaches are only just gaining momentum. For us, it is very exciting to help shape the methods that could soon be used all over the world.”

Publication

Georg Sperl, Rosa M. Sánchez-Banderas, Manwen Li, Chris Wojtan und Miguel A. Otaduy. 2022. Estimation of Yarn-Level Simulation Models for Production Fabrics. ACM Transactions on Graphics, Vol. 41, No. 4, Art. 65 (Proceedings of SIGGRAPH). DOI: 10.1145/3528223.3530167

Funding

This research project was partially funded by the European Research Council (ERC) Consolidator Grant 772738 TouchDesign.