Reproduce “RealMe: Immersive Volumetric Video from Minimal Captures” 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

Chenxu Zhou, Zhiyuan Yu, Yudong Jin, Yinji Shentu, Yu Zhang, Jiaming Sun, Xiaowei Zhou

RealMe improves free-viewpoint video from sparse, low-resolution captures by generating coordinated views, enhancing them with a diffusion model that uses time and camera layout, and reconstructing a 3D scene. On RealVV, with eight input views, it achieves a novel-view PSNR of 27.2706 versus 24.4387 for Diffuman4D. This result applies to the paper’s 2560×1440 evaluation protocol.

Summary by alphaXiv.

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