Reproduce “All-St3R SLAM: Real-time RGB Dynamic SLAM with Pixel-wise Motion Masking” 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
Yewon Kim, Paul Hongsuck Seo, Sang Chul Ahn
All-St3R SLAM filters moving pixels during camera mapping by comparing motion predicted from camera movement with image-space motion tracked by All-Tracker, rather than masking whole object categories. On Bonn, it reports 10.13 FPS and 2.46 cm ATE RMSE, versus WildGS-SLAM (Fast) at 2.13 FPS and 2.47 cm. Its results depend on adequate geometric separation and can fail.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.