Reproduce “Pixel-wise Planarity for High-Precision Monocular Plane Segmentation” 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
Ahmetcan Yavuz, Alpay Ozkan, Rémi Pautrat, Shaohui Liu, Marc Pollefeys
The method segments planes in single images by predicting pixel-wise planarity confidence alongside depth and surface normals, then grouping pixels whose geometric cues agree. On ScanNet++, it reaches 62.1% 3D plane precision at 1 mm, compared with 16.6% for retrained ZeroPlane. This comparison retains different model formulations and does not isolate the planarity head alone.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.