Reproduce “Accord: Universal Online Metric Localization and Dense Mapping from one Camera and an IMU” 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
Jaafar Mahmoud, Arthur Movsesyan, Cuong Huynh, Mikhail Iumanov, Sergey Zagoruyko, Sergey Kolyubin
Accord builds dense metric maps by combining sparse visual-inertial odometry points with learned depth. It anchors depth windows to metric points, then compares overlapping views to reconcile image-scale differences while preserving the anchor’s mean scale. Across four public scan-backed datasets, it has the highest mean 5-centimeter F-score among evaluated dense pipelines; this does not establish absolute scale independently of VIO.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.