Abstract

We present FMPlug, a novel plug-in framework that enhances foundationflow-matching (FM) priors for solving ill-posed inverse problems. Unliketraditional approaches that rely on domain-specific or untrained priors, FMPlugsmartly leverages two simple but powerful insights: the similarity betweenobserved and desired objects and the Gaussianity of generative flows. Byintroducing a time-adaptive warm-up strategy and sharp Gaussianityregularization, FMPlug unlocks the true potential of domain-agnostic foundationmodels. Our method beats state-of-the-art methods that use foundation FM priorsby significant margins, on image super-resolution and Gaussian deblurring.

 

Quick Read (beta)