pith. sign in

arxiv: 2212.00490 · v2 · pith:XQEYLUXJnew · submitted 2022-12-01 · 💻 cs.CV

Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model

classification 💻 cs.CV
keywords ddnmrestorationdiffusionimagemodelnull-spacezero-shotdenoising
0
0 comments X
read the original abstract

Most existing Image Restoration (IR) models are task-specific, which can not be generalized to different degradation operators. In this work, we propose the Denoising Diffusion Null-Space Model (DDNM), a novel zero-shot framework for arbitrary linear IR problems, including but not limited to image super-resolution, colorization, inpainting, compressed sensing, and deblurring. DDNM only needs a pre-trained off-the-shelf diffusion model as the generative prior, without any extra training or network modifications. By refining only the null-space contents during the reverse diffusion process, we can yield diverse results satisfying both data consistency and realness. We further propose an enhanced and robust version, dubbed DDNM+, to support noisy restoration and improve restoration quality for hard tasks. Our experiments on several IR tasks reveal that DDNM outperforms other state-of-the-art zero-shot IR methods. We also demonstrate that DDNM+ can solve complex real-world applications, e.g., old photo restoration.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 22 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. PolarVSR: A Unified Framework and Benchmark for Continuous Space-Time Polarization Video Reconstruction

    cs.CV 2026-05 unverdicted novelty 8.0

    PolarVSR is the first unified architecture for continuous space-time polarization video reconstruction from DoFP captures, using polarization-aware implicit neural representations, a flow-guided variation loss, and a ...

  2. FlowBender: Feedback-Aware Training for Self-Correcting Conditional Flows

    cs.CV 2026-06 unverdicted novelty 7.0

    FlowBender introduces closed-loop training that lets conditional flow models learn correction policies from their own task-specific alignment errors, outperforming supervised and guidance baselines on fidelity and pla...

  3. Reflection Separation from a Single Image via Joint Latent Diffusion

    cs.CV 2026-06 unverdicted novelty 7.0

    A joint latent diffusion model with cross-layer self-attention and disjoint sampling separates reflection and transmission layers from single images more effectively than prior methods on real-world benchmarks.

  4. Proximal-Based Generative Modeling for Bayesian Inverse Problems

    math.OC 2026-05 unverdicted novelty 7.0

    PGM replaces the intractable likelihood score in diffusion models with a closed-form Moreau score computed via proximal operators, enabling non-asymptotic sampling for inverse problems trained only on prior data.

  5. Proximal-Based Generative Modeling for Bayesian Inverse Problems

    math.OC 2026-05 unverdicted novelty 7.0

    PGM framework links diffusion to proximal regularization for closed-form Moreau-score sampling in Bayesian inverse problems, learned only from prior samples.

  6. Invertible Diffusion for Low-Memory Channel Gain Map Construction in Wireless Communication Networks

    eess.SP 2026-04 unverdicted novelty 7.0

    InvDiff-CGM uses invertible architectures in diffusion and U-Net plus a multi-scale prior injector to construct CGMs with 85% lower peak training memory and 38.02 dB PSNR on RadioMap3DSeer.

  7. Diff-ANO: Towards Fast High-Resolution Ultrasound Computed Tomography via Conditional Consistency Models and Adjoint Neural Operators

    math.NA 2025-07 unverdicted novelty 7.0

    Diff-ANO uses conditional consistency models and adjoint neural operator surrogates to enable fast, high-quality USCT reconstructions under sparse and partial views by replacing slow PDE solvers and enabling few-step ...

  8. COCO-Inpaint: A Benchmark for Detecting and Localizing Inpainting-Based Image Manipulations

    cs.CV 2025-04 unverdicted novelty 7.0

    COCO-Inpaint supplies a large-scale dataset and evaluation protocol focused on inpainting-based image forgeries to benchmark existing detection methods.

  9. From Sparse X-rays to 3D CT: Training-Free Reconstruction with Diffusion Priors

    eess.IV 2026-06 unverdicted novelty 6.0

    TF-PRDiT uses a frozen 3D diffusion transformer prior with task-specific forward operators and predictor-corrector sampling to solve X-ray-to-CT reconstruction and other volumetric inverse problems without any retraining.

  10. Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance

    cs.CV 2026-06 unverdicted novelty 6.0

    Moebius introduces a compressed diffusion inpainting model using Local-λ Mix Interaction blocks and latent-space multi-granularity distillation to reach 10B-level quality with 0.22B parameters.

  11. Measurement Geometry and Design for Trustworthy Generative Inverse Problems

    cs.LG 2026-06 unverdicted novelty 6.0

    Introduces a local measurement-manifold compatibility measure that bounds stable reconstruction error under generative priors and motivates fixed and adaptive acquisition rules.

  12. Conservative Flows: A New Paradigm of Generative Models

    cs.LG 2026-05 unverdicted novelty 6.0

    Conservative flows generate by running probability-preserving stochastic dynamics initialized at data points rather than noise, using corrected Langevin or predictor-corrector mechanisms on top of any pretrained flow ...

  13. Plug-and-Play Label Map Diffusion for Universal Goal-Oriented Navigation

    cs.RO 2026-05 unverdicted novelty 6.0

    PLMD applies a denoising diffusion model to predict labels for unknown map regions, allowing goal localization in unexplored environments by substituting completed labels into existing navigation pipelines.

  14. Conditional Diffusion Under Linear Constraints: Langevin Mixing and Information-Theoretic Guarantees

    cs.LG 2026-05 unverdicted novelty 6.0

    Error in approximating the tangent conditional score by the unconditional score in diffusion models is bounded by dimension-free conditional mutual information, with a projected-Langevin method outperforming baselines...

  15. Degradation-Robust Fusion: An Efficient Degradation-Aware Diffusion Framework for Multimodal Image Fusion in Arbitrary Degradation Scenarios

    cs.CV 2026-04 unverdicted novelty 6.0

    A degradation-aware diffusion framework fuses multimodal images under arbitrary degradations by directly regressing the fused image and applying joint degradation-fusion constraints during limited-step sampling.

  16. Analyzing and Guiding Zero-Shot Posterior Sampling in Diffusion Models

    cs.LG 2026-02 unverdicted novelty 6.0

    Under a Gaussian prior assumption, zero-shot diffusion posterior samplers for inverse problems admit closed-form spectral representations that enable a new parameter-selection framework balancing perceptual quality an...

  17. Protein Autoregressive Modeling via Multiscale Structure Generation

    cs.LG 2026-02 unverdicted novelty 6.0

    PAR is a multi-scale autoregressive transformer framework for protein backbone generation that uses coarse-to-fine prediction, noisy context learning, and flow-based decoding to achieve high-quality unconditional and ...

  18. NPN: Non-Linear Projections of the Null-Space for Imaging Inverse Problems

    cs.CV 2025-10 unverdicted novelty 6.0

    NPN introduces a neural-network-based regularization that promotes reconstructions lying in a low-dimensional projection of the sensing operator's null-space, with claimed theoretical guarantees and improved empirical...

  19. Separating Intrinsic Ambiguity from Estimation Uncertainty in Deep Generative Models for Linear Inverse Problems

    cs.LG 2026-05 unverdicted novelty 5.0

    Introduces a cascade-based structural decomposition of posterior uncertainty to isolate intrinsic ambiguity from estimation uncertainty in deep generative models for linear inverse problems.

  20. Unifying Deep Stochastic Processes for Image Enhancement

    cs.CV 2026-05 unverdicted novelty 5.0

    Stochastic image enhancement methods are shown to be variants of a shared SDE differing in drift, diffusion, terminal distributions and boundary conditions, with controlled experiments revealing no single dominant fam...

  21. Dual Ascent Diffusion for Inverse Problems

    cs.CV 2025-05 unverdicted novelty 5.0

    A dual ascent optimization framework is introduced for MAP estimation with diffusion priors, claimed to outperform prior methods on image restoration in quality, noise robustness, speed, and data fidelity.

  22. A Survey on Diffusion Models for Inverse Problems

    cs.LG 2024-09 unverdicted novelty 5.0

    A survey that introduces taxonomies for categorizing pre-trained diffusion model methods applied to inverse problems and analyzes their connections and challenges.