Reproduce “Prior-Guided Single-Exposure HDR Novel View Synthesis via Base-Gain Representation” 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
Armin Shafiee Sarvestani, Sheyang Tang, Jialu Xu, Zhou Wang
PriorGS enables novel HDR and LDR view synthesis from single-exposure LDR views. It uses pretrained 2D HDR estimates to guide a 3D Gaussian scene with an LDR base, gain field, and learned camera response function for re-exposure. On synthetic HDR-NeRF, it achieves 31.22 dB HDR-view PSNR, versus 26.96 dB for GaussHDR; real scenes lack HDR ground truth.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.