Reproduce “Spatons: Spatiotonally Localized Building Blocks for Efficient XR Predistortion” 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
Mohammadjavad Maheronnaghsh, Sushant Kondguli, Steve Mann, Andreas Moshovos
Spatons decomposes XR image predistortion into local geometric resampling and color refinement. Sequential-5′ uses pixel coordinates to predict sampling locations, then a patch-based network to adjust local color. In modeled Scenario 5, it reaches SSIM 0.944 and PSNR 27.5 dB, matching CompenNet++ on SSIM with 6× fewer FLOPs and 106× fewer parameters. Tests are mostly modeled, without validation through XR hardware.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.