Reproduce “CroCoMAE: Cross-modal self-supervised collaborative representation learning for SAR-optical remote sensing data” 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
Vito Recchia, Giuseppe Guarino, Cássio Fraga Dantas, Giuseppina Andresini, Annalisa Appice, Dino Ienco
CroCoMAE learns representations from paired SAR and optical images for land-cover mapping. It aligns shared dimensions of corresponding spatial patches and reconstructs masked inputs for each sensor. On DFC2020, with 1% labels and a frozen encoder, it scores 51.4% mean intersection-over-union on SAR, versus 48.0% for its no-partition variant, supporting the design in the tested setting, not universal superiority.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.