Delaunay Canopy (ECCV 2026): SOTA building wireframe reconstruction from sparse LiDAR
RexDouglass · x · 2026-09-05
Researchers from Yonsei and Emory introduce Delaunay Canopy, an ECCV 2026 paper reconstructing building wireframes from sparse, noisy airborne LiDAR point clouds (poster Fri Sep 11).
- Core idea: Delaunay triangulation of the point cloud, with curvature signatures computed at face, edge, and vertex levels as geometric priors to score the Delaunay graph
- Corner selection uses vertex-wise priors to isolate likely true corners, then refines them; wire selection scores candidates along shortest Delaunay paths to adaptively shrink the search space
- State-of-the-art on Building3D Tallinn city and entry-level datasets, with comprehensive experiments
More from Research
- Randomized YaRN: training on short context with sampled positions boosts 128K reasoning — gregd_nlp · 2026-09-05
- OpenAI Details New Training Methods for Non-Verifiable Domains in Astra — morqon · 2026-09-05
- Richard Szeliski's Computer Vision textbook, 2nd ed., free PDF with 210k+ downloads — alexbilz · 2026-09-05
- Procedural Generation and Modern ML Share the Same Ethos: Sampling Conditioned Distributions — keenanisalive · 2026-09-05
- John Schulman: this research is timely as CoT monitorability declines — johnschulman2 · 2026-09-05
- Princeton team says OpenAI's new architecture closely resembles its T2MLR paper — prfsanjeevarora · 2026-09-05