Reproduce “DrawSplat: Atlanta World-Guided Implicit Gaussian SLAM for Indoor Reconstruction” 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
Fan Zhu, Zhisong Xu, Ziyu Chen, Zhenjun Zhao, Takayuki Okatani
DrawSplat separates indoor SLAM into structural tracking and mapping. It estimates camera motion and architectural planes from tracked points and lines, then uses an anchor-based implicit Gaussian map to add appearance and local geometry. On Replica Apartment, it reports 1.09-centimeter trajectory-error RMSE; on ScanNet, mean PSNR is 27.07 dB. These results concern RGB-D indoor scenes with dominant architectural structure.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.