Reproduce “GaussFlaw: Fine-Grained Perceptual Diagnosis for 3D Gaussian 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
Shaocheng Shen, Zihan Zheng, Kejin Li, Lu Liu, Xiaoyun Zhang, Qiang Hu, Wenjun Zhang
GeoCritic combines RGB appearance and depth-derived geometry to localize and classify rendering flaws, while GaussFlaw-80K provides 80,049 labeled regions in 14,289 images. It reaches AUC-J 0.9507 and Acc@1 80.08%. Diagnosis-guided refinement improves global and diagnosed-region crop metrics across the tested datasets and backbones, with crop metrics computed over all pixels within fixed crop boundaries.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.