Reproduce “Learned Smartphone ISP on Mobile GPUs: Mobile AI 2026 Challenge Report” 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
Andrey Ignatov, Georgy Perevozchikov, Radu Timofte, Yuhong He, Yuxuan Qiu, Tianyu Hao, Ruoqi Li, Yueqi Yang and 14 more
MiAlgo maps Bayer RAW to RGB photos with a learned ISP designed for mobile hardware. Its LKSNet processes 1920×2560-pixel images evaluated on two flagship phone GPUs. On the final test set, it achieved PSNR of 24.19 dB, SSIM of 0.8587, and MOS of 4.5, with average GPU runtime of 27.5 ms. These results apply to tested devices.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.