Reproduce “DLOSAC: Joint Spatial-Geometric Image Estimation via Delaunay-Guided Local Optimization RANSAC” 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
Yu Liu, Meiqi Zhou, Zipei Fan, Guanyuan Feng, Xuan Song
DLOSAC refines geometric estimates from RANSAC inliers by combining point-level residuals with edge-consistency and local Delaunay-triangle distribution cues in weighted least squares. These spatial cues enter the same point weights used for refinement. Across four fundamental-matrix datasets, mean pixel error averages 0.156 px, versus 0.179 px for PSC-RANSAC. Gains are limited on wide-baseline datasets; DLOSAC is not a RANSAC replacement.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.