Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

PnP-Flow: Plug-and-Play Image Restoration with Flow Matching

About

In this paper, we introduce Plug-and-Play (PnP) Flow Matching, an algorithm for solving imaging inverse problems. PnP methods leverage the strength of pre-trained denoisers, often deep neural networks, by integrating them in optimization schemes. While they achieve state-of-the-art performance on various inverse problems in imaging, PnP approaches face inherent limitations on more generative tasks like inpainting. On the other hand, generative models such as Flow Matching pushed the boundary in image sampling yet lack a clear method for efficient use in image restoration. We propose to combine the PnP framework with Flow Matching (FM) by defining a time-dependent denoiser using a pre-trained FM model. Our algorithm alternates between gradient descent steps on the data-fidelity term, reprojections onto the learned FM path, and denoising. Notably, our method is computationally efficient and memory-friendly, as it avoids backpropagation through ODEs and trace computations. We evaluate its performance on denoising, super-resolution, deblurring, and inpainting tasks, demonstrating superior results compared to existing PnP algorithms and Flow Matching based state-of-the-art methods.

S\'egol\`ene Martin, Anne Gagneux, Paul Hagemann, Gabriele Steidl• 2024

Related benchmarks

TaskDatasetResultRank
Image InpaintingFFHQ (test)
LPIPS0.234
73
Super-Resolution (4x)ImageNet
PSNR24.4
57
Motion DeblurFFHQ
PSNR27.1
56
Multi-Coil MRI ReconstructionfastMRI Brain 8x acceleration multi-coil
SSIM0.913
37
Multi-Coil MRI ReconstructionfastMRI Brain multi-coil 4x acceleration
SSIM95.1
37
Super-ResolutionFFHQ (test)
SSIM44.2
32
HDRImageNet
PSNR17.8
31
DenoisingAFHQ Cat (test)
PSNR31.76
30
Super-ResolutionCelebA (test)
PSNR31.49
30
Phase RetrievalImageNet
PSNR15.2
29
Showing 10 of 71 rows
...

Other info

Follow for update