Reproduce “Spaghetti Splatting: Differentiating Curvilinear Dirac Masses via Convolution” 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
George Shramko, Shrisudhan Govindarajan, Rudy Lam, Yizheng Xie, Daniel Rebain, David B. Lindell, Andrea Tagliasacchi
Spaghetti Splatting reconstructs scenes from images using connected polylines whose segments are convolved with Gaussian or Beta kernels, then rendered through volume compositing. This extends point splats to continuous, strand-like geometry while retaining image-based optimization. On selected hair benchmarks, it reaches an F1 score of 61.2 versus 46.9 for Gaussian Haircut at 2 mm and 20 degrees, without establishing topology.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.