| Department |
Knitting is one of the most popular techniques to manufacture clothing. In this dissertation we present a framework for modeling, fabricating, and simulating knit structures. We first introduce a fully automatic pipeline to convert arbitrary 3D shapes into a stitch mesh, which provides a mesh-based representation of the yarn-level knit structure. We also extend the concept of stitch mesh modeling with a list of novel shaping techniques to allow modeling more complex structures. We then introduce a scheduling algorithm to convert the structure into step-by-step hand knitting instructions for knitters. We further embed low-level machine knitting operations to each face and allow users to edit the stitch mesh in a way that preserves the machine knittability. Moreover, we present a real-time rendering method to display knitwear containing more than a hundred million individual fiber curves at real-time frame rates with shadows and ambient occlusion on the GPU. We also present a GPU parallelization method to address the computational challenges of using the Material Point Method to simulate knits. |
