Reproduce “Perception Over Geometry: Rethinking Gaussian Allocation in Mesh-based Splatting” 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
Yuting Wang, Yang Zhao, Shanding Diao, Penglin Wang, Yuan Chen, Wei Jia, Ronggang Wang
P-MGS allocates Gaussian splats according to estimated appearance demand rather than mesh density, then binds freely optimized splats to a reference mesh for editing. It estimates demand from visible coverage and image-gradient energy, while mesh deformation updates Gaussian centers and covariances. On NeRF-Synthetic, P-MGS reports 34.00 dB average PSNR with 545K Gaussians, versus 33.25 dB and 2,687K for Mani-GS, a benchmark-specific result.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.