Reproduce “WildFab: Multi-Axis 3D Printing from Models in the Wild” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jiasheng Qu, Zhikai Shen, Chenyu Xu, Hailin Sun, Chengkai Dai, Yuhu Guo, Junpeng Wang, Yeung Yam and 1 more
WildFab generates multi-axis toolpaths and plans collisions from open surfaces and mixed thin shells and solids, without converting them into watertight solids. It combines a neural unsigned distance field with a regularized generalized winding number field to locate surfaces and guide planning. Collision checking is reported 73× faster than reg-GWN-only checking, but sharp or small-scale features may be smoothed.
Key insights
- A hybrid geometric representation, combining the differentiable distance properties of neural UDFs with the robust solid-void semantics and surface ridge detection of regularized GWNs, effectively addresses the challenges of "models in the wild."
- The gradient magnitude ridge of a regularized GWN provides a precise and robust mechanism for surface localization, even for open or non-manifold geometries, when guided by UDF gradients.
- Differentiable queries from the hybrid field enable efficient optimization for fabricating complex geometries, facilitating simultaneous support-free toolpath generation and collision-free robot motion planning.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.