Reproduce “Online Segment 3D Gaussians via Launching Virtual Drones” 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
Liwei Liao, Rongjie Wang, Ronggang Wang
SAGO tackles interactive object selection in raw 3DGS scenes without per-scene preparation. It uses virtual cameras to collect masks from new viewpoints, which progressively remove background Gaussians. On NVOS, SAGO reports 92.7% mIoU and 98.7% mAcc in 0.4–0.9 seconds, compared with GaussianCut’s 92.5%, 98.4%, and 40–60 seconds. These are benchmark segmentation times; prompt-mask time is excluded, and the method requires a high-quality initial view.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.