Reproduce “Pano2Room: Novel View Synthesis from a Single Indoor Panorama” 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

Guo Pu, Yiming Zhao, Zhouhui Lian

Pano2Room turns one room panorama into a textured mesh, fills unseen regions iteratively with RGB-D inpainting, and trains a 3D Gaussian Splatting representation on generated views. It selects cameras to target missing content. On Replica’s Office 0, it scores 25.15 PSNR versus PERF’s 21.15, performing best on most scenes. Hidden content is generated, not verified as the room’s true appearance.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.