Reproduce “Transform your Smartphone into a DSLR Camera: Learning the ISP in the Wild” 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
Ardhendu Shekhar Tripathi, Martin Danelljan, Samarth Shukla, Radu Timofte, Luc Van Gool
The authors learn a phone RAW-to-sRGB pipeline for DSLR-style rendering by separating color prediction from detail restoration. During training, a forward–backward optical-flow consistency mask excludes regions with unreliable alignment. On ISPW, the method reports 25.05 dB PSNR and 0.821 SSIM, versus LiteISPNet’s 23.51 dB and 0.809. Evidence covers two benchmarks, not performance across phones or scenes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.