Reproduce “BEV-GS: Feed-forward Gaussian Splatting in Bird's-Eye-View for Road Reconstruction” 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
Wenhua Wu, Tong Zhao, Chensheng Peng, Lei Yang, Yintao Wei, Zhe Liu, Hesheng Wang
BEV-GS predicts a road height map and appearance from one frame, using a bird’s-eye-view grid and elevation-guided texture features to form Gaussians for rendering another view. On the RSRD test set, it reports 1.73 cm elevation AEE, versus 1.93 cm for RoadBEV, and 28.36 dB PSNR, versus 23.27 dB for Flash3D. These results concern this dataset and setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.