Reproduce “SLGaussian: Fast Language Gaussian Splatting in Sparse Views” 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
Kangjie Chen, BingQuan Dai, Minghan Qin, Dongbin Zhang, Peihao Li, Yingshuang Zou, Haoqian Wang
SLGaussian predicts a queryable semantic 3D Gaussian scene from two RGB views, avoiding scene-specific optimization at test time. It links consistent object masks to low-dimensional IDs and stores multi-view CLIP features for language queries. On four LERF scenes, it achieves 41.50% overall queried-object IoU and 70.80% localization accuracy, above LangSplat’s overall scores, though results vary by scene and metric.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.