Reproduce “Gaussian Light Field Splatting: A Physical Prior-Driven Vision Transformer for Low-Light Image Enhancement” 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
Yuhan Chen, Wenxuan Yu, Guofa Li, Fuchen Li, Kunyang Huang, Yicui Shi, Ying Fang, Wenbo Chu and 1 more
GLFS enhances low-light images by embedding Gaussian-splatting illumination geometry in Transformer attention, guiding feature interactions and spatial brightness correction. Its tokenizer models overlapping, directionally shaped light regions, while a decoder produces the enhanced image. On LOL, full GLFS reaches 24.38 dB PSNR and 0.84 SSIM, versus 23.92 dB and 0.79 without Gaussian attention; results do not establish zero-shot performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.