Pith. sign in

REVIEW 4 major objections 8 minor 66 references

Navigating Image Restoration with VAR's Distribution Alignment Prior

T0 review · 4 major / 8 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A frozen visual autoregressive model's multi-scale predictions supply a clean-image distribution alignment prior, and a restoration network built on it, VarFormer, outperforms existing multi-task restoration methods on six degradation…

desk verdict A competent all-in-one restoration paper with a released codebase whose headline claim about VAR's 'endogenous' prior is not yet supported, because the training pipeline itself manufactures much of the alignment it attributes to frozen VAR. read the letter →

arxiv 2412.21063 v2 pith:4W4WNUTH submitted 2024-12-30 cs.CV

Siyang Wang , Feng Zhao This is my paper · ORCID
classification cs.CV
keywords imagerestorationvisualautoregressivemodelgenerativepriormulti-scalerepresentationall-in-onedistributionalignmentzero-shotgeneralizationtransformerarchitecture
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that a frozen visual autoregressive model, VAR, carries a reusable clean-image prior in its multi-scale latent representations: as generation advances scale by scale, the model's next-scale predictions automatically modulate degraded input representations toward the clean-image distribution, with early scales handling global degradations such as haze and low light and later scales handling local degradations such as noise and rain. On the strength of that observation, the authors build VarFormer, a single all-in-one restoration network that first extracts these scale-wise prior features with a frozen VAR plus a lightweight adapter, then feeds them into a U-shaped restoration backbone through degradation-aware weighting and adaptive feature transformation. If the central claim is right, a generative model trained only for image synthesis can supply a general-purpose prior that restores many degradation types in one model while cutting training cost, and the same prior transfers to unseen degradations without fine-tuning.

What carries the argument

The load-bearing object is the frozen VAR transformer's next-scale prediction, the autoregressive factorization $p(r_1,\dots,r_K)=\prod_{k=1}^K p(r_k \mid r_1,\dots,r_{k-1})$ in which a sequence of token maps of increasing resolution is predicted, each conditioned on the previous scales. The paper treats the collection of VAR's predicted scale features along with the VQVAE encoder features as a multi-scale clean-image prior, and argues that the alignment with clean images emerges as the autoregressive predictions modulate the input scale by scale. Two training stages carry the argument: Stage 1 freezes VAR and trains a cross-attention probe on clean images, followed by an adapter with a feature-matching loss to reduce the degraded-clean latent gap; Stage 2 uses Degradation-Aware Enhancement modules, which compute region-specific fusion weights over re-weighted scale priors through Swin-Transformer blocks and a projection layer, and Adaptive Feature Transformation modules, which perform two-step attention with a low-dimensional mediator, to guide a U-shaped restoration backbone. An adaptive mix-up skip merges encoder features into the decoder to limit detail loss. The machinery's job is to convert an architectural byproduct of generation, scale-wise distribution alignment, into a controllable restoration signal.

What would settle it

Repeat the replacement experiment of Eq. (2) with the VAR transformer's weights replaced by a randomly initialized transformer of the same architecture, keeping the clean-image-trained cross-attention probe and the VQVAE decoder fixed; if scale-specific degradation removal (haze at low scales, noise at high scales) still appears at similar strength, the effect is not a learned property of VAR's autoregressive distribution. Separately, train the Stage-1 adapter only on clean images and measure a per-scale distribution distance between encoder features and VAR-predicted features for clean versus degraded inputs; if no scale shows a reduced gap, the central observation is contradicted.

Watch

Extended reading notes

Core claim

VarFormer's central discovery is that the multi-scale token maps produced by VAR's next-scale autoregressive prediction contain an endogenous distribution-alignment prior: when degraded image latents are passed through the frozen VAR, replacing certain encoder-derived scale features with the model's predicted scale features removes specific degradations—low scales remedy global degradations such as low light and haze, high scales remedy local degradations such as Gaussian noise and rain. t-SNE visualizations of the latent representations indicate that VAR's next-scale prediction narrows the gap between degraded and clean latent distributions. The paper operationalizes this prior in two stages: Stage 1 freezes VAR, trains a cross-attention probe on clean images only, and then adds an adapter with a feature-matching loss to close the residual domain gap between degraded and clean latents; Stage 2 injects the resulting multi-scale embeddings into a restoration network through Degradation-Aware Enhancement modules, which predict region-specific fusion weights over re-weighted priors, and Adaptive Feature Transformation modules, which mediate attention through a low-dimensional bridge feature. Trained once on six tasks, the model outperforms existing multi-task restoration methods on deraining, deblurring, dehazing, low-light enhancement, Gaussian denoising, and real denoising, and it transfers zero-shot to under-display-camera restoration.

Load-bearing premise

The alignment between degraded and clean representations is an intrinsic property of the frozen pretrained VAR's next-scale predictions, not something manufactured by the paper's own clean-image reconstruction probe or its adapter and feature-matching loss; if those trained modules are doing most of the aligning, the conceptual framing collapses even though the empirical restoration results may still stand.

Editorial extensions

If this is right

  • A single VarFormer model trained once on six degradations restores all of them, and on low-light enhancement and dehazing it even surpasses task-specific models such as Restormer that were trained for a single task.
  • The frozen VAR prior transfers to an unseen degradation, under-display-camera artifacts, without any fine-tuning, improving PSNR, SSIM, or LPIPS over prior universal methods on the TOLED and POLED benchmarks.
  • Because VAR stays frozen and only the adapter and restoration modules are trained, the pipeline reduces training computational cost relative to training a restoration model from scratch with generative guidance.
  • Ablation results attribute the gain to each component: removing the adapter, the adaptive skip, the AFT module, or the DAE module lowers average PSNR/SSIM across the six tasks, with the full model reaching 29.66 dB average PSNR.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The scale-to-degradation correspondence (low scales for global effects, high scales for local effects) suggests a diagnostic use: the replacement experiment of Eq. (2) could localize which scales a given corruption attacks, potentially informing blind restoration or degradation classification.
  • The conceptual claim depends on the prior being intrinsic to the frozen VAR, so an immediate test is to train the Stage-1 cross-attention probe on degraded-clean pairs instead of clean images only, or to replace it with a linear probe; if the alignment largely vanishes, much of the 'endogenous' prior is actually learned by the training signal.
  • Because the extraction mechanism is generic, other vector-quantized autoregressive image models with hierarchical latents could be tested for the same degradation-specific scale alignment, which would tell whether the effect is specific to VAR or shared across next-scale generative paradigms.
  • The zero-shot UDC gains likely reflect that under-display degradations are a blend of global tone loss and local blur; a stress test on fundamentally different artifacts, such as JPEG compression or super-resolution aliasing, would bound how general the 'unseen task' claim really is.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 8 minor

Summary. The paper proposes VarFormer, a unified image-restoration framework that exploits multi-scale latent representations from a frozen Visual Autoregressive (VAR) model as a generative prior. The authors report an observation that during VAR-based reconstruction, replacing degraded-image encoder features with VAR-transformer-predicted features at selected scales removes different degradation types, and they interpret this as evidence of an "endogenous" multi-scale distribution alignment capability. They then build a two-stage pipeline: Stage 1 trains an Adapter and a feature-matching loss to produce multi-scale distribution-aligned embeddings from the frozen VAR, and Stage 2 fuses these embeddings into a restoration U-Net through degradation-aware enhancement and adaptive feature transformation modules. The method is evaluated on deraining, deblurring, dehazing, low-light enhancement, Gaussian denoising, real-image denoising, and zero-shot under-display-camera restoration, with code released.

Significance. If the claimed endogenous distribution alignment prior were established, it would be a novel and potentially reusable clean-image prior for multi-task restoration, with possible training-cost advantages from freezing the generative backbone. The paper provides a fairly broad empirical evaluation and releases code. However, the central conceptual claim is currently not cleanly separated from the paper's own training signals: the observation in Sec. 3.1 relies on a cross-attention probe trained only on clean images, and Stage 1 explicitly trains an adapter with a feature-matching loss to align degraded features with clean ones. The significance of the work therefore depends on control experiments that isolate a pre-existing property of the frozen VAR from the effects of the learned probe and adapter.

major comments (4)
  1. [Sec. 3.1 and Eq. (2)] The observation of an "endogenous" distribution alignment prior is confounded by the clean-image-trained cross-attention probe. The reconstruction training in Sec. 3.1 is conducted only on clean images, and the cross-attention is the only trainable module; all other modules are frozen. When degraded images are later fed through this pipeline, the VAR Transformer's predictions F_deg^v are conditioned on encoder features that have passed through a probe trained to map into the clean-image reconstruction distribution. The apparent scale-wise degradation removal could therefore be an artifact of the probe rather than a property of the pretrained VAR itself. Please provide a control in which degraded images are processed by the frozen VAR without the clean-trained cross-attention probe, and show whether the scale-replacement experiment in Eq. (2) still removes degradations. Without this control, the claim that VAR "naturally" or "automatically" aligns inputs with clean-image distributions is unsupported.
  2. [Sec. 3.2 and Eq. (7)] The Stage-1 training explicitly manufactures the distribution alignment. The feature-matching loss ||F_a - sg(F_q^egt)||_2^2 in Eq. (7) trains the Adapter to push degraded encoder features toward clean VQVAE features, and the text states this is done "to further enhance the adaptive distribution alignment capability" of VAR. The "multi-scale distribution alignment prior" fed to Stage 2 is therefore not an unchanged property of the frozen VAR but a learned embedding produced by the paper's own training signal. Please either (a) ablate the feature-matching loss and the cross-entropy term in Eq. (7) and report restoration performance with the frozen VAR features alone, or (b) reframe the contribution as a trained alignment adapter combined with a frozen VAR prior. The current framing overstates the novelty of the prior.
  3. [Sec. 4.3 and Table 4] The statement that VarFormer "achieves state-of-the-art performance" on unknown tasks is not consistent with the reported numbers. On POLED, VarFormer's PSNR (16.63 dB) is lower than IDR's (16.71 dB), and its LPIPS (0.605) is considerably worse than DiffUIR's (0.505). On TOLED, the PSNR gain over DiffUIR is 1.06 dB but the SSIM is identical (0.887). Please qualify the zero-shot claim, discuss the POLED discrepancy, or provide additional analysis such as a per-image breakdown or statistical significance tests.
  4. [Tables 1, 2, and 4] Several margins over the closest competing universal method are small, for example deraining 31.33 vs 31.14 dB, low-light enhancement 25.13 vs 25.02 dB, dehazing 32.96 vs 32.74 dB, and many Gaussian-denoising entries within 0.1-0.3 dB. No error bars, multiple-seed results, or statistical tests are reported, so the claim of "superior performance across all 6 tasks" is not yet established at the reported precision. Please provide variance estimates or significance tests, or temper the claim accordingly.
minor comments (8)
  1. [Title and Abstract] The paper consistently types "V AR" with a space; use "VAR" throughout.
  2. [Abstract] The sentence "Extensive experiments underscores that our VarFormer outperforms..." has a subject-verb agreement error; it should be "underscore."
  3. [Sec. 3.1] The phrase "we need it to model specific images" is unclear, and "we provide them to V AR" has an ambiguous antecedent. Consider rewriting for clarity.
  4. [Sec. 3.2] The text says "the V AR that has been finetuned for reconstruction in Sec. 3.1," but in Sec. 3.1 only the cross-attention module was trained; the VAR itself was frozen. Please correct this description.
  5. [Eq. (7)] The notation "sg" and "F_q^egt" is not defined in the text, and the loss label "Lf ema" is rendered with an odd spacing. Please define these symbols and use a consistent notation.
  6. [Sec. 4.1] The sentence "We random crop256×256 patch from original image as network input after data augmentation" is missing a space and an article; it should be "We randomly crop 256×256 patches from the original image after data augmentation."
  7. [Sec. 4.1] The task-sampling weights (0.3 for dehazing, 0.1 for low-light, etc.) are given without justification; consider reporting how sensitive the final performance is to these weights.
  8. [Fig. 3] The t-SNE figure mentions "w/o VAR reconstruction" and "with VAR reconstruction," but the manuscript does not define what "w/o VAR reconstruction" means in this context. Please clarify what is being compared.

Circularity Check

2 steps flagged · score 6.0 of 10

The 'endogenous distribution alignment prior' is partly manufactured by the paper's own clean-trained cross-attention probe and Stage-1 feature-matching loss, so the central conceptual claim is partially circular; the external benchmark gains are not.

  1. fitted input called prediction [Sec. 3.1 / Fig. 2, Eq. (2)]
    "we employ cross-attention to inject multi-scale embeddings Fe obtained from the VQVAE encoder into the scale autoregression Transformer in VAR [40] for image reconstruction pre-training. During this process, we freeze the VAR and train only the cross-attention mechanism using clean images. ... Following that, we replace the clean images with degraded images to obtain the GT index sequence F deg_e ... and the sequence F deg_v predicted by the Transformer."

    The F^deg_v features used in Eq. (2) to demonstrate that degradation disappears when scale features are replaced are produced by a trainable cross-attention module that was optimized on clean images. Feeding degraded encoder features into this clean-trained probe biases the Transformer's predictions toward the clean-image manifold, so the observed 'automatic' distribution alignment and degradation removal can be an artifact of the probe's training distribution rather than an endogenous property of frozen VAR. The paper presents this as evidence of VAR's 'adaptive distribution alignment capability,' but the capability is partly supplied by the fitted probe.

  2. self definitional [Sec. 3.2 'Adapter for Domain Shift' / Sec. 3.3 Eq. (7)]
    "we freeze the VAR and integrate an Adapter to deliberately reduce the distance between the multi-scale latent representations of clean and degraded images, thereby obtaining multi-scale distribution alignment embedding S_v. ... The goal of the first stage is to further enhance the distribution alignment capability of VAR. Consequently, we design the Feature Matching Loss ... L_fema = sum_i -s_i log(s_hat_i) + ||F_a - sg(F_q^egt)||_2^2."

    The feature-matching term ||F_a - sg(F_q^egt)||_2^2 directly optimizes the Adapter output F_a toward the clean quantized features. The 'multi-scale distribution alignment embedding S_v' used by Stage 2 and later attributed to VAR's 'endogenous multi-scale priors' is therefore produced by the paper's own training objective, not discovered in the frozen VAR. The claim that VAR 'adaptively aligns the input with clean images scale by scale' reduces by construction to the loss the authors themselves impose, so the conceptual novelty is circular even though the restoration numbers remain externally valid.

full rationale

The empirical restoration results in Tables 1-4 are benchmarked against external methods and use an externally pretrained VAR, so the performance claims are not self-referential or circular. However, the paper's headline conceptual contribution—that frozen VAR possesses an 'endogenous' multi-scale distribution alignment prior—is not cleanly separated from the paper's own training signals. The Fig. 2 observation relies on a cross-attention module trained on clean images, and the Stage-1 Adapter is explicitly trained with a feature-matching loss that minimizes the distance between degraded and clean features. The extracted 'prior' is then presented as a pre-existing property of VAR. Since no frozen-VAR-only control is reported, the central conceptual claim is partially circular: the alignment evidence is manufactured by the training objective. This does not invalidate the benchmark gains, but it does mean the attribution of those gains to a discovered 'endogenous' prior is unsupported. Score 6 reflects this partial, load-bearing circularity in the conceptual derivation, not in the external evaluation.

Assumptions & free parameters 2 free parameters · 4 assumptions · 1 invented entities

The framework contributes neither a formal derivation nor new measured constants. The load-bearing content is a qualitative observation about a pretrained model, plus hand-set task weights. The central assumptions concern the clean-image fidelity of VAR and the scale-degradation coupling.

free parameters (2)
  • task_sampling_weights = 0.2 deraining, 0.3 dehazing, 0.1 low-light, 0.2 Gaussian denoising, 0.1 real denoising, 0.1 motion deblurring
    Hand-set in Sec. 4.1 to balance the six tasks in each training batch. These weights directly control the all-in-one trade-off and no sensitivity analysis is reported.
  • initial_learning_rate = 1e-4 decaying to 1e-6
    Chosen schedule for Adam with cosine annealing. It is a standard choice, but the convergence claim in Sec. 1 is not backed by an experiment comparing schedules or measuring training time.
assumptions (4)
  • domain assumption The pretrained VAR model reliably captures the distribution of clean natural images through its scale-wise VQVAE and autoregressive transformer.
    This is the basis of the entire prior. It is inherited from the VAR paper [40] and not re-verified on restoration data here.
  • domain assumption Degradation type is coupled to scale: global degradations are removed by low-scale features and local ones by high-scale features.
    Derived from the scale-replacement experiments in Fig. 2 with 100 paired images per degradation. The paper uses this coupling to justify the DAE weighting in Eq. (3).
  • domain assumption t-SNE separation between degraded and clean features is a valid indicator of distribution alignment.
    Fig. 3 shows only t-SNE projections; no scalar divergence metric such as FID or MMD is computed, so the strength of the claimed alignment is not measured.
  • ad hoc to paper Replacing encoder scale features with VAR-transformer-predicted scale features in Eq. (2) removes degradations, making those features useful restoration priors.
    This replacement probe is the paper's central observation, but it uses a cross-attention module trained only on clean images, so it does not isolate the frozen VAR alone.
invented entities (1)
  • Multi-scale distribution alignment prior
    purpose: Conceptual claim that VAR's scale predictions inherently align degraded inputs with the clean-image distribution, used to justify the DAE and AFT modules.
    No external falsifiable prediction is made beyond the paper's own benchmark results. The prior is evidenced only by the paper's qualitative t-SNE and scale-replacement experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Navigating Image Restoration with VAR's Distribution Alignment Prior." pith.science (2026). https://pith.science/paper/4W4WNUTH

@misc{pith2026241221063,
  author       = {Pith},
  title        = {Pith review of: Navigating Image Restoration with VAR's Distribution Alignment Prior},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4W4WNUTH}},
  note         = {Machine review of arXiv:2412.21063}
}
read the original abstract

Generative models trained on extensive high-quality datasets effectively capture the structural and statistical properties of clean images, rendering them powerful priors for transforming degraded features into clean ones in image restoration. VAR, a novel image generative paradigm, surpasses diffusion models in generation quality by applying a next-scale prediction approach. It progressively captures both global structures and fine-grained details through the autoregressive process, consistent with the multi-scale restoration principle widely acknowledged in the restoration community. Furthermore, we observe that during the image reconstruction process utilizing VAR, scale predictions automatically modulate the input, facilitating the alignment of representations at subsequent scales with the distribution of clean images. To harness VAR's adaptive distribution alignment capability in image restoration tasks, we formulate the multi-scale latent representations within VAR as the restoration prior, thus advancing our delicately designed VarFormer framework. The strategic application of these priors enables our VarFormer to achieve remarkable generalization on unseen tasks while also reducing training computational costs. Extensive experiments underscores that our VarFormer outperforms existing multi-task image restoration methods across various restoration tasks.

Figures

Figures reproduced from arXiv: 2412.21063 by the authors.

Figure 1
Figure 1. Motivation of VarFormer. (1) As the autoregressive scale [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of our investigation about the multi-scale distribution alignment priors within VAR. (a) Reconstruction Training: To [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. The t-SNE diagrams demonstrate that VAR’s next-scale [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: The framework of our VarFormer includes two training stages. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Visual comparison with state-of-the-art methods on image deraining task. Please zoom in for details. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Visual comparison with state-of-the-art methods on low-light image enhancement task. Please zoom in for details. [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Visualization of weight maps from DAE [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 47 canonical work pages

  1. [1]

    Im- age2stylegan: How to embed images into the stylegan latent space? In Proceedings of the IEEE/CVF international con- ference on computer vision, pages 4432–4441, 2019

    Rameen Abdal, Yipeng Qin, and Peter Wonka. Im- age2stylegan: How to embed images into the stylegan latent space? In Proceedings of the IEEE/CVF international con- ference on computer vision, pages 4432–4441, 2019. 3

    work page 2019

  2. [2]

    A high-quality denoising dataset for smartphone cameras

    Abdelrahman Abdelhamed, Stephen Lin, and Michael S Brown. A high-quality denoising dataset for smartphone cameras. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 1692– 1700, 2018. 6

    work page 2018

  3. [3]

    Derin Babacan, Rafael Molina, and Aggelos K

    S. Derin Babacan, Rafael Molina, and Aggelos K. Katsagge- los. Variational bayesian blind deconvolution using a total variation prior. IEEE Transactions on Image Processing, 18: 12–26, 2009. 3

    work page 2009

  4. [4]

    Large scale gan training for high fidelity natural image synthesis

    Andrew Brock. Large scale gan training for high fidelity natural image synthesis. arXiv preprint arXiv:1809.11096 ,

    arXiv

  5. [5]

    Hinet: Half instance normalization network for image restoration

    Liangyu Chen, Xin Lu, Jie Zhang, Xiaojie Chu, and Cheng- peng Chen. Hinet: Half instance normalization network for image restoration. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 182–192, 2021. 7, 8

    work page 2021

  6. [6]

    Simple baselines for image restoration

    Liangyu Chen, Xiaojie Chu, Xiangyu Zhang, and Jian Sun. Simple baselines for image restoration. In Proceedings of the European Conference on Computer Vision, pages 17–33,

    work page

  7. [7]

    Sam2-adapter: Evaluating & adapting seg- ment anything 2 in downstream tasks: Camouflage, shadow, medical image segmentation, and more

    Tianrun Chen, Ankang Lu, Lanyun Zhu, Chaotao Ding, Chu- nan Yu, Deyi Ji, Zejian Li, Lingyun Sun, Papa Mao, and Ying Zang. Sam2-adapter: Evaluating & adapting seg- ment anything 2 in downstream tasks: Camouflage, shadow, medical image segmentation, and more. arXiv preprint arXiv:2408.04579, 2024. 1

    arXiv 2024

  8. [8]

    Challenges on large scale surveillance video analysis

    Weitao Feng, Deyi Ji, Yiru Wang, Shuorong Chang, Han- sheng Ren, and Weihao Gan. Challenges on large scale surveillance video analysis. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 69–76, 2018. 1

    work page 2018

Show all 66 references
  1. [9]

    Kodak lossless true color image suite

    Rich Franzen. Kodak lossless true color image suite. http: //r0k.us/graphics/kodak , 1999. Accessed: 1999- 06-07. 6

    1999

  2. [10]

    Clearing the skies: A deep network architecture for single-image rain removal

    Xueyang Fu, Jiabin Huang, Xinghao Ding, Yinghao Liao, and John William Paisley. Clearing the skies: A deep network architecture for single-image rain removal. IEEE Transactions on Image Processing, 26:2944–2956, 2016. 6

    2016

  3. [11]

    Removing rain from single images via a deep detail network

    Xueyang Fu, Jiabin Huang, Delu Zeng, Yue Huang, Xing- hao Ding, and John William Paisley. Removing rain from single images via a deep detail network. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1715–1723, 2017. 6

    2017

  4. [12]

    Image processing using multi-code gan prior

    Jinjin Gu, Yujun Shen, and Bolei Zhou. Image processing using multi-code gan prior. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 3012–3021, 2020. 3

    2020

  5. [13]

    Zero-reference deep curve estimation for low-light image enhancement

    Chunle Guo, Chongyi Li, Jichang Guo, Chen Change Loy, Junhui Hou, Sam Kwong, and Runmin Cong. Zero-reference deep curve estimation for low-light image enhancement. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 1780–1789, 2020. 1

    2020

  6. [14]

    Single image haze removal using dark channel prior.IEEE Transactions on Pat- tern Analysis and Machine Intelligence , 33(12):2341–2353,

    Kaiming He, Jian Sun, and Xiaoou Tang. Single image haze removal using dark channel prior.IEEE Transactions on Pat- tern Analysis and Machine Intelligence , 33(12):2341–2353,

  7. [15]

    Single image super-resolution from transformed self-exemplars

    Jia-Bin Huang, Abhishek Singh, and Narendra Ahuja. Single image super-resolution from transformed self-exemplars. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 5197–5206, 2015. 6

    2015

  8. [16]

    Ultra-high resolution segmentation with ultra-rich con- text: A novel benchmark

    Deyi Ji, Feng Zhao, Hongtao Lu, Mingyuan Tao, and Jieping Ye. Ultra-high resolution segmentation with ultra-rich con- text: A novel benchmark. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 23621–23630, 2023. 1

    2023

  9. [17]

    Progressive growing of gans for improved qual- ity, stability, and variation.arXiv preprint arXiv:1710.10196,

    Tero Karras. Progressive growing of gans for improved qual- ity, stability, and variation.arXiv preprint arXiv:1710.10196,

    arXiv

  10. [18]

    A style-based generator architecture for generative adversarial networks

    Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4401–4410, 2019

    2019

  11. [19]

    Analyzing and improv- ing the image quality of stylegan

    Tero Karras, Samuli Laine, Miika Aittala, Janne Hellsten, Jaakko Lehtinen, and Timo Aila. Analyzing and improv- ing the image quality of stylegan. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8110–8119, 2020. 3

    2020

  12. [20]

    Denoising diffusion restoration models

    Bahjat Kawar, Michael Elad, Stefano Ermon, and Jiaming Song. Denoising diffusion restoration models. Advances in Neural Information Processing Systems, 35:23593–23606,

  13. [21]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 ,

    arXiv

  14. [22]

    Benchmarking single- image dehazing and beyond

    Boyi Li, Wenqi Ren, Dengpan Fu, Dacheng Tao, Dan Feng, Wenjun Zeng, and Zhangyang Wang. Benchmarking single- image dehazing and beyond. IEEE Transactions on Image Processing, 28(1):492–505, 2018. 6

    2018

  15. [23]

    All-in-one image restoration for unknown corruption

    Boyun Li, Xiao Liu, Peng Hu, Zhongqin Wu, Jiancheng Lv, and Xiaocui Peng. All-in-one image restoration for unknown corruption. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17431– 17441, 2022. 1, 7, 8

    2022

  16. [24]

    Recurrent squeeze-and-excitation context ag- gregation net for single image deraining

    Xia Li, Jianlong Wu, Zhouchen Lin, Hong Liu, and Hong- bin Zha. Recurrent squeeze-and-excitation context ag- gregation net for single image deraining. arXiv preprint arXiv:1807.05698, 2018. 6

    2018 arXiv

  17. [25]

    Tan, Xiaojie Guo, Jiangbo Lu, and M

    Yu Li, Robby T. Tan, Xiaojie Guo, Jiangbo Lu, and M. S. Brown. Rain streak removal using layer priors. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2736–2744, 2016. 6

    2016

  18. [26]

    Swinir: Image restoration us- ing swin transformer

    Jingyun Liang, Jiezhang Cao, Guolei Sun, Kai Zhang, Luc Van Gool, and Radu Timofte. Swinir: Image restoration us- ing swin transformer. InProceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 1833–1844,

  19. [27]

    Diff- bir: Towards blind image restoration with generative diffu- sion prior

    Xinqi Lin, Jingwen He, Ziyan Chen, Zhaoyang Lyu, Bo Dai, Fanghua Yu, Wanli Ouyang, Yu Qiao, and Chao Dong. Diff- bir: Towards blind image restoration with generative diffu- sion prior. arXiv preprint arXiv:2308.15070, 2023. 2, 3

    2023 arXiv

  20. [28]

    Residual denoising diffu- sion models

    Jiawei Liu, Qiang Wang, Huijie Fan, Yinong Wang, Yan- dong Tang, and Liangqiong Qu. Residual denoising diffu- sion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 2773– 2783, 2024. 7

    2024

  21. [29]

    Tape: Task-agnostic prior embedding for image restoration

    Lin Liu, Lingxi Xie, Xiaopeng Zhang, Shanxin Yuan, Xi- angyu Chen, Wengang Zhou, Houqiang Li, and Qi Tian. Tape: Task-agnostic prior embedding for image restoration. In Proceedings of the European Conference on Computer Vi- sion, pages 447–464, 2022. 8

    2022

  22. [30]

    Hcanet: Haze-concentration-aware network for real- scene dehazing with codebook priors

    Yi Liu, Jiachen Li, Yanchun Ma, Qing Xie, and Yongjian Liu. Hcanet: Haze-concentration-aware network for real- scene dehazing with codebook priors. In Proceedings of the 32nd ACM International Conference on Multimedia , pages 9136–9144, 2024. 4

    2024

  23. [31]

    Removing rain from a sin- gle image via discriminative sparse coding

    Yu Luo, Yong Xu, and Hui Ji. Removing rain from a sin- gle image via discriminative sparse coding. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 3397–3405, 2015. 6

    2015

  24. [32]

    Controlling vision-language models for universal image restoration

    Ziwei Luo, Fredrik K Gustafsson, Zheng Zhao, Jens Sj¨olund, and Thomas B Sch¨on. Controlling vision-language models for universal image restoration. arXiv preprint arXiv:2310.01018, 2023. 7, 8

    2023 arXiv

  25. [33]

    Waterloo ex- ploration database: New challenges for image quality assess- ment models

    Kede Ma, Zhengfang Duanmu, Qingbo Wu, Zhou Wang, Hongwei Yong, Hongliang Li, and Lei Zhang. Waterloo ex- ploration database: New challenges for image quality assess- ment models. IEEE Transactions on Image Processing , 26 (2):1004–1016, 2016. 6

    2016

  26. [34]

    A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics

    David Martin, Charless Fowlkes, Doron Tal, and Jitendra Malik. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics. In Proceedings of the IEEE/CVF International Conference on Computer Vision...

    2001

  27. [35]

    Deep generalized unfolding networks for image restoration

    Chong Mou, Qian Wang, and Jian Zhang. Deep generalized unfolding networks for image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17399–17410, 2022. 3, 8

    2022

  28. [36]

    Deep multi-scale convolutional neural network for dynamic scene deblurring

    Seungjun Nah, Tae Hyun Kim, and Kyoung Mu Lee. Deep multi-scale convolutional neural network for dynamic scene deblurring. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 257–265,

  29. [37]

    Promp- tir: Prompting for all-in-one blind image restoration

    V Potlapalli, SW Zamir, S Khan, and FS Khan. Promp- tir: Prompting for all-in-one blind image restoration. arXiv preprint arXiv:2306.13090, 2023. 7

    2023 arXiv

  30. [38]

    Neumann network with recursive kernels for single image defocus deblurring

    Yuhui Quan, Zicong Wu, and Hui Ji. Neumann network with recursive kernels for single image defocus deblurring. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5754–5763, 2023. 1

    2023

  31. [39]

    Learning to deblur using light field generated and real de- focus images

    Lingyan Ruan, Bin Chen, Jizhou Li, and Miuling Lam. Learning to deblur using light field generated and real de- focus images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16304– 16313, 2022. 1

    2022

  32. [40]

    Visual autoregressive modeling: Scalable im- age generation via next-scale prediction

    Keyu Tian, Yi Jiang, Zehuan Yuan, Bingyue Peng, and Li- wei Wang. Visual autoregressive modeling: Scalable im- age generation via next-scale prediction. arXiv preprint arXiv:2404.02905, 2024. 2, 3

    2024 arXiv

  33. [41]

    Stripformer: Strip transformer for fast image deblurring

    Fu-Jen Tsai, Yan-Tsung Peng, Yen-Yu Lin, Chung-Chi Tsai, and Chia-Wen Lin. Stripformer: Strip transformer for fast image deblurring. In Proceedings of the European Confer- ence on Computer Vision, pages 146–162, 2022. 1

    2022

  34. [42]

    Maxim: Multi-axis mlp for image processing

    Zhengzhong Tu, Hossein Talebi, Han Zhang, Feng Yang, Peyman Milanfar, Alan Bovik, and Yinxiao Li. Maxim: Multi-axis mlp for image processing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5769–5780, 2022. 7

    2022

  35. [43]

    Transweather: Transformer-based restoration of im- ages degraded by adverse weather conditions

    Jeya Maria Jose Valanarasu, Rajeev Yasarla, and Vishal M Patel. Transweather: Transformer-based restoration of im- ages degraded by adverse weather conditions. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2353–2363, 2022. 3

    2022

  36. [44]

    Exploiting diffusion prior for real-world image super-resolution

    Jianyi Wang, Zongsheng Yue, Shangchen Zhou, Kelvin CK Chan, and Chen Change Loy. Exploiting diffusion prior for real-world image super-resolution. arXiv preprint arXiv:2305.07015, 2023. 3

    2023 arXiv

  37. [45]

    Images speak in images: A generalist painter for in-context visual learning

    Xinlong Wang, Wen Wang, Yue Cao, Chunhua Shen, and Tiejun Huang. Images speak in images: A generalist painter for in-context visual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6830–6839, 2023. 7

    2023

  38. [46]

    Zero-shot im- age restoration using denoising diffusion null-space model

    Yinhuai Wang, Jiwen Yu, and Jian Zhang. Zero-shot im- age restoration using denoising diffusion null-space model. arXiv preprint arXiv:2212.00490, 2022. 2, 3

    2022 arXiv

  39. [47]

    Uformer: A general u-shaped transformer for image restoration

    Zhendong Wang, Xiaodong Cun, Jianmin Bao, and Jianzhuang Liu. Uformer: A general u-shaped transformer for image restoration. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 17662–17672, 2021. 7

    2021

  40. [48]

    Lg-bpn: Local and global blind-patch network for self-supervised real-world denoising

    Zichun Wang, Ying Fu, Ji Liu, and Yulun Zhang. Lg-bpn: Local and global blind-patch network for self-supervised real-world denoising. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 18156–18165, 2023. 1

    2023

  41. [49]

    Deep retinex decomposition for low-light enhancement

    Chen Wei, Wenjing Wang, Wenhan Yang, and Jiaying Liu. Deep retinex decomposition for low-light enhancement. arXiv preprint arXiv:1808.04560, 2018. 6

    2018 arXiv

  42. [50]

    Learning semantic-aware knowledge guidance for low-light image enhancement

    Yuhui Wu, Chen Pan, Guoqing Wang, Yang Yang, Jiwei Wei, Chongyi Li, and Heng Tao Shen. Learning semantic-aware knowledge guidance for low-light image enhancement. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 1662–1671, 2023. 1

    2023

  43. [51]

    Low-light image enhancement via structure modeling and guidance

    Xiaogang Xu, Ruixing Wang, and Jiangbo Lu. Low-light image enhancement via structure modeling and guidance. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 9893–9903, 2023. 1

    2023

  44. [52]

    Tan, Jiashi Feng, Jiaying Liu, Zong- ming Guo, and Shuicheng Yan

    Wenhan Yang, Robby T. Tan, Jiashi Feng, Jiaying Liu, Zong- ming Guo, and Shuicheng Yan. Deep joint rain detec- tion and removal from a single image. In Proceedings of 10 the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1685–1694, 2016. 6

    2016

  45. [53]

    Tan, Shiqi Wang, Yuming Fang, and Jiaying Liu

    Wenhan Yang, Robby T. Tan, Shiqi Wang, Yuming Fang, and Jiaying Liu. Single image deraining: From model-based to data-driven and beyond. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43:4059–4077, 2019. 6

    2019

  46. [54]

    Ad- verse weather removal with codebook priors

    Tian Ye, Sixiang Chen, Jinbin Bai, Jun Shi, Chenghao Xue, Jingxia Jiang, Junjie Yin, Erkang Chen, and Yun Liu. Ad- verse weather removal with codebook priors. InProceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 12653–12664, 2023. 4

    2023

  47. [55]

    Learning enriched features for real image restoration and enhancement

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, and Ling Shao. Learning enriched features for real image restoration and enhancement. In Proceedings of the European Confer- ence on Computer Vision, pages 492–511, 2020. 7, 8

    2020

  48. [56]

    Multi-stage progressive image restoration

    Syed Waqas Zamir, Aditya Arora, Salman Hameed Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, and Ling Shao. Multi-stage progressive image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14816–14826, 2021. 1, 7, 8

    2021

  49. [57]

    Restormer: Efficient transformer for high-resolution image restoration

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Mu- nawar Hayat, Fahad Shahbaz Khan, and Ming-Hsuan Yang. Restormer: Efficient transformer for high-resolution image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 5728– 573...

    2022

  50. [58]

    He Zhang and Vishal M. Patel. Density-aware single image de-raining using a multi-stream dense network. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 695–704, 2018. 6

    2018

  51. [59]

    Sindagi, and Vishal M

    He Zhang, Vishwanath A. Sindagi, and Vishal M. Patel. Im- age de-raining using a conditional generative adversarial net- work. IEEE Transactions on Circuits and Systems for Video Technology, 30:3943–3956, 2017. 6

    2017

  52. [60]

    Accurate image restora- tion with attention retractable transformer

    Jiale Zhang, Yulun Zhang, Jinjin Gu, Yongbing Zhang, Linghe Kong, and Xin Yuan. Accurate image restora- tion with attention retractable transformer. arXiv preprint arXiv:2210.01427, 2022. 7

    2022 arXiv

  53. [61]

    Ingredient-oriented multi- degradation learning for image restoration

    Jinghao Zhang, Jie Huang, Mingde Yao, Zizheng Yang, Hu Yu, Man Zhou, and Feng Zhao. Ingredient-oriented multi- degradation learning for image restoration. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5825–5835, 2023. 1, 3, 7, 8

    2023

  54. [62]

    Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising

    Kai Zhang, Wangmeng Zuo, Yunjin Chen, Deyu Meng, and Lei Zhang. Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising. IEEE Transactions on Image Processing, 26(7):3142–3155, 2017. 1

    2017

  55. [63]

    Prac- tical blind denoising via swin-conv-unet and data synthesis

    Kai Zhang, Yawei Li, Jingyun Liang, Jiezhang Cao, Yulun Zhang, Hao Tang, Radu Timofte, and Luc Van Gool. Prac- tical blind denoising via swin-conv-unet and data synthesis. arXiv preprint arXiv:2203.13278, 2022. 1

    2022 arXiv

  56. [64]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 586–595, 2018. 6

    2018

  57. [65]

    Selective hourglass mapping for universal image restoration based on diffusion model

    Dian Zheng, Xiao-Ming Wu, Shuzhou Yang, Jian Zhang, Jian-Fang Hu, and Wei-Shi Zheng. Selective hourglass mapping for universal image restoration based on diffusion model. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 25445– 25455,...

    2024

  58. [66]

    Image restoration for under-display camera

    Yuqian Zhou, David Ren, Neil Emerton, Sehoon Lim, and Timothy Large. Image restoration for under-display camera. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9179–9188, 2021. 6 11

    2021

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.