IR-Flow uses rectified flow to build linear transport between degraded and clean images via multilevel flows and cumulative velocity fields, enabling competitive restoration quality with only a few sampling steps.
For ODE solver, we choose a first-order numerical solver, as shown in Algorithm 1 and 2
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
IR-Flow: Bridging Discriminative and Generative Image Restoration via Rectified Flow
IR-Flow uses rectified flow to build linear transport between degraded and clean images via multilevel flows and cumulative velocity fields, enabling competitive restoration quality with only a few sampling steps.