ReFlowSET is a conditional latent flow-matching framework for converting synthetic aperture radar (SAR) imagery into electro-optical (EO) imagery. Rather than relying on a pretrained autoencoder, it selects an optimal codec through joint SAR-EO reconstruction evaluation, then trains a smaller conditional diffusion transformer from scratch using dual-stream SAR conditioning and representation alignment with frozen vision foundation models. The approach achieves state-of-the-art results on the QXS-SAROPT and SAR2Opt benchmarks.
