Pith. sign in

REVIEW 4 major objections 5 minor 50 references

Towards a Universal Image Degradation Model via Content-Degradation Disentanglement

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper claims a single learned module can encode, separate, and transfer arbitrary image degradations, including spatially varying ones, without user parameters.

desk verdict A genuinely new degradation-transfer architecture with an overclaimed theoretical guarantee; worth serious review, but the universality claim should be dialed back. read the letter →

arxiv 2505.12860 v2 pith:4AA5H5IM submitted 2025-05-19 cs.CV eess.IV

Wenbo Yang , Zhongling Wang , Zhou Wang This is my paper · ORCID
classification cs.CVeess.IV
keywords imagedegradationsynthesistransfercontent-degradationdisentanglementhomogeneousinhomogeneousblindrestorationfilmgrainsimulationentropyregularization
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

This paper aims to establish that a single learned model can encode the degradation present in any distorted image, separate it from the image content, and reapply that degradation to any other image without user-supplied degradation parameters. The claim matters because image-restoration pipelines, film-grain simulation, and artistic effects currently rely on narrow, per-distortion models with hand-tuned parameters, which fail on the complex and spatially varying distortions common in real photos. If the claim holds, one module can replace many specialized degradation simulators and can turn non-blind restoration systems into blind ones.

What carries the argument

The core mechanism is disentangle-by-compression: entropy regularization applied to the homogeneous and inhomogeneous degradation embeddings e_g and e_l. The rate losses sum the per-entry entropies, and the identity sum_i H(e^(i)) = H(e) + D_KL(p(e)||q(e)) means that minimizing the loss both reduces the total entropy and forces the entries toward independence. The homogeneous encoding network uses a dual-branch design with short- and long-range receptive fields, while the inhomogeneous encoding network keeps spatial structure; the synthesis network inserts a new deconvolution-based layer, IDA, combined with a spatial feature transform in an IDA-SFT block to apply spatially varying and global degradations.

What would settle it

Train the same architecture on a paired dataset where the applied degradation is deliberately content-dependent (for example, blur strength proportional to local saliency or haze depth derived from scene content), then measure whether the homogeneous embedding can be used to decode or predict the clean image content, or whether degradation transfer to a different-content image loses fidelity. If content is recoverable from the embedding or transfer quality drops, the central disentanglement claim is falsified.

Watch

Extended reading notes

Core claim

The central claim is that degradation and content are disentanglable by compression: a pair of encoding networks extracts a global, spatially uniform degradation embedding and a local, spatially structured embedding from a distorted image, and a synthesis network rebuilds the degradation onto any clean image. Training with an entropy regularization loss on both embeddings makes the embeddings carry as little information as possible. Under the paper's stated assumptions, minimizing the entropy yields the identity H(e)=I(e;x)+H(d), so the embedding's mutual information with the clean content is driven toward zero while the fixed entropy of the degradation remains, separating the two. The same loss also pushes embedding entries toward statistical independence, giving interpretable latent dimensions that control groups of degradations. With this mechanism, the paper reports high reproduction and transfer scores on synthetic and real distortions, successful film-grain transfer, and blind restoration with two inversion-based methods.

Load-bearing premise

The disentanglement guarantee assumes the distortion process is statistically independent of the clean image content and that spatially uniform and spatially varying degradations are independent of each other; real-world degradations are often content-dependent, so when that assumption fails the proof no longer ensures the embedding is free of content.

Editorial extensions

If this is right

  • A single pre-trained degradation module could replace per-distortion simulators in training-data generation for restoration, super-resolution, and denoising tasks.
  • Non-blind inversion-based restoration methods that plug in this module become blind: they only need the distorted image, not degradation parameters, as demonstrated on two distinct inversion-based frameworks.
  • Film-grain encoding and transfer become a single operation: the model can take a grainy frame, encode its grain, and apply it to a grain-free frame while preserving content, matching or exceeding a specialized grain method's reproduction score.
  • The degradation embeddings are interpretable: perturbing one active dimension changes one family of degradations, so degradation can be edited as a controllable attribute rather than a fixed pipeline step.

Reading between the lines

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

  • Because the model is trained only on paired clean/distorted images, its generality is bounded by the diversity of that training data; a natural testable extension is to train on unpaired multi-distortion corpora and measure whether transfer accuracy scales with degradation diversity rather than content diversity.
  • If the entropy-loss derivation holds, an analogous disentangle-by-compression objective might separate other spatially varying attributes from content, such as lighting, albedo, or style, wherever the same independence assumptions approximately hold.
  • The model's blindness relies on the test-time distortion being represented by its training distribution; one could probe the boundary by feeding it a degradation outside the training pool and checking whether the embedding silently reinterprets it as a familiar distortion rather than failing openly.
  • A practical consequence of interpretable embedding dimensions is a potential editing interface: users could dial a single latent dimension to strengthen or weaken a specific distortion without retraining, something the paper demonstrates visually but does not develop as a tool.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The manuscript proposes a universal image degradation model that encodes a degraded image into a homogeneous degradation embedding e_g and a spatially varying embedding e_l using two encoder networks (HDEN and IDEN), then synthesizes degradations on a target pristine image through a U-Net with IDA-SFT blocks. The central methodological claim is a “disentangle-by-compression” approach: an entropy regularization loss on the embeddings is shown, under assumptions A-1 to A-5, to yield H(e) = I(e;x) + H(d), so that minimizing H(e) minimizes the mutual information between the degradation embedding and image content. The paper reports experiments on synthetic degradation reproduction and transfer, film grain synthesis, and the conversion of non-blind inversion-based restoration methods (RSG, DPS) into blind ones.

Significance. If the disentanglement guarantee holds, the paper would be a valuable step toward learning degradation representations that can be transferred across images without user-provided degradation parameters, and the idea of plugging such a module into inversion-based restoration to make it blind is practically interesting. The paper includes informative ablations showing that the entropy loss and the IDA/IDEN components improve transfer scores, and it promises code and data. However, the “universal” claim is broader than the theoretical and experimental support: the central identity is derived under assumptions that are in tension with content-dependent degradations, and the experiments cover a limited set of distortion types without direct comparisons to existing degradation models.

major comments (4)
  1. [Sec. 3.3, Eq. (6), Assumptions A-2 and A-5] The derivation of H(e) = I(e;x) + H(d) requires the step H(e|x) = H(d) with H(d) a constant independent of x, which follows from A-2 (degradation distribution independent of the clean image) and, for the homogeneous/inhomogeneous split, A-5 (joint independence of homogeneous and inhomogeneous degradations). The paper's own introduction states that real-world complex degradations are typically nonlinear and content-dependent. Depth-dependent haze, object-motion blur, and rain all violate these assumptions: the degradation parameters include scene depth, object motion, or coupled global/local rain effects, which are correlated with content. If A-2 fails, minimizing H(e) may suppress content-related information rather than isolate a content-free degradation representation, which would directly undermine the transfer experiments of Sec. 4.1 and the blind-restoration claim of Sec. 4.4. Please state the theorem with all assumptions explicitly and provide experiments on content-dependent degradations (e.g., depth-based haze, object-dependent motion blur) to show the disentanglement still holds or to delimit the failure regime.
  2. [Sec. 3.3, proof of Eq. (6)] The central proof is deferred to Supp. Sec. 6.2, and the main text does not specify the probability space: which variables are random, over which set of images and distortions the entropies are taken, and how the random state n is treated. This makes Eq. (6) impossible to verify from the main text. At minimum, state the theorem and proof sketch in the main text, or move the proof into the main text, because Eq. (6) is the theoretical justification for the entire disentangle-by-compression loss.
  3. [Sec. 4.4, Eq. (9), Table 4] The blind-restoration evaluation compares only “w/o Ours” (naive blind inversion without re-degradation) with “w/ Ours”. There is no comparison with the original non-blind method that uses the true degradation parameters, nor with other blind restoration baselines. The claim that the model achieves “competitive performance” and converts non-blind restoration into blind restoration needs a calibration point: for example, report the non-blind upper bound from [30] on the same test set, and report the performance of a blind method that estimates degradation parameters directly.
  4. [Sec. 4.1 and Sec. 4.5, Tables 2–6] The “universal” claim is supported only by a restricted synthetic pool (typical image-processing pipeline degradations on WQIs), film grain, and raindrop transfer. No quantitative comparison is provided with existing degradation models, including the closest prior work [12], despite Table 1 positioning the paper against them. Please add comparisons on at least the degradations used by previous models; if no comparison is possible, explicitly state that limitation. Otherwise “first universal degradation model” is not supported by the current evidence.
minor comments (5)
  1. [Eq. (7)] In the total loss, λ_g is used for both Lrate_g and L_gan, which is ambiguous. Use a distinct symbol such as λ_gan for the adversarial loss weight.
  2. [Table 4] Several entries are run together without spacing, e.g., “228.828.1” and “45.116.9”. Please format the table so the values are readable.
  3. [Sec. 3, first sentence; Fig. 5 caption; Table 1 header] There are several typos: “all distorted process can be described” should be “all distorted processes can be described”; “Encodeing” in Fig. 5 should be “Encoding”; “Mdl Type” in Table 1 should be “Model Type”.
  4. [Sec. 3.3, entropy estimation] The text says p(e_g) cannot be reliably estimated “due to its high dimensionality,” but the loss uses estimates of p(e_g(i)); please clarify exactly which density is estimated and how the high-dimensionality issue is avoided.
  5. [Sec. 4.1] The sentence “Since this is the first universal image degradation model, it is difficult to find comparable models” is an explanation, not a substitute for comparison; if comparable methods are not available, the claim should be softened accordingly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the disentanglement guarantee is a self-contained information-theoretic derivation under stated assumptions, with held-out evaluations.

full rationale

The derivation chain is self-contained. The disentanglement guarantee H(e)=I(e;x)+H(d) (Eq. 6) is a mathematical consequence of the stated assumptions A-1..A-5, proved in Supp. Sec. 6.2; it is not obtained by fitting a parameter to the result it is used to explain. The entropy regularization loss (Sec. 3.3) is a regularizer, and the evaluation of reproduction/transfer uses held-out WQI test triplets (Sec. 4.1), the FilmGrainStyle test set (Sec. 4.3), and the FFHQ-X restoration benchmarks (Sec. 4.4) that the model did not train on. Self-citations ([37], [38]) are non-load-bearing: SSIM is used as a diversity regularizer, and [38] only supports a post-hoc interpretation of latent dimensions. The main caveat is that A-2 and A-5 (degradation independent of content; homogeneous/inhomogeneous degradations jointly independent) conflict with the paper's own observation that real degradations are 'typically nonlinear and content-dependent'; if these assumptions fail, Eq. (6) and the disentanglement guarantee may not hold. This is a correctness/robustness limitation, not a circularity: the paper does not assume the target result, and no fitted parameter is renamed as a prediction.

Assumptions & free parameters 3 free parameters · 8 assumptions · 4 invented entities

The central claim rests on a set of explicitly stated assumptions (A-1 to A-5) that connect the entropy regularization loss to information-theoretic disentanglement. The loss trade-off weights and a few architectural dimensions are free parameters tuned by hand. The new network modules are internally validated but lack independent evidence outside the paper.

free parameters (3)
  • Loss trade-off weights (λ_g, λ_l, λ_c, λ_r, λ_d, λ_gan) = not reported in main text
    Weights in the total loss (Eq. 7) are chosen by hand/validation; they control the balance between disentanglement, reconstruction fidelity, diversity, and realism. The proof of disentanglement in Section 3.3 requires 'mild conditions' on λ_g and λ_l.
  • Downsampling factor d in HDEN long-range branch = not reported
    Hyperparameter in the long-range branch that operates on a downsampled resolution (Section 3.1).
  • Embedding dimensions of e_g and e_l = not reported
    Dimensionality of the degradation embeddings; affects capacity and the number of active dimensions observed (Section 4.2).
assumptions (8)
  • standard math Sum-entropy identity Σ_i H(e_i) = H(e) + D_KL(p(e)∥∏_i p(e_i))
    Used in Eq. (5) to decompose the regularization loss into total entropy and independence terms.
  • domain assumption A-1: Given degradation process and clean image, the degraded image can be exactly reconstructed up to the random state
    Assumption used to derive H(e) = I(e;x) + H(d) in Section 3.3.
  • domain assumption A-2: Distribution of distortion processes is independent of the clean images
    Load-bearing assumption in the disentanglement proof; contradicted by content-dependent degradations the paper itself mentions in the introduction.
  • domain assumption A-3: Distortion can be inferred from the degraded image
    Ensures the embedding e contains the degradation information; used in the proof.
  • domain assumption A-4: With an appropriate perceptual loss L_sim, the reconstructed degraded image is similar enough to the target
    Justifies using DISTS as the reconstruction loss and enables the entropy argument.
  • domain assumption A-5: Homogeneous and inhomogeneous degradations are jointly independent
    Used to show the optimal solution of the weighted entropy loss separates e_l and e_g; unlikely to hold for coupled real distortions like haze with spatially varying density.
  • domain assumption The model is sufficiently flexible to achieve the optimum of the entropy-regularized loss
    The proof that e_l only contains spatially dependent information requires 'flexible enough network' and 'mild conditions' on λ's (Section 3.3).
  • domain assumption DISTS is an appropriate perceptual distance with low sensitivity to noise random state
    The choice of loss determines what information the model is allowed to discard; stated in Section 3.4.
invented entities (4)
  • Homogeneous Degradation Embedding Network (HDEN)
    purpose: Encodes spatial-invariant degradation information e_g from a distorted image.
    A new architectural module introduced in Section 3.1; its effectiveness is shown only through internal ablations in the paper.
  • Inhomogeneous Degradation Embedding Network (IDEN)
    purpose: Encodes spatially varying degradation information e_l while retaining spatial structure.
    Introduced in Section 3.1; evaluated only within the paper's own experiments.
  • Inhomogeneous Degradation Aware (IDA) layer
    purpose: Provides a computationally efficient way to apply spatially varying degradation during synthesis.
    Defined in Eq. (3); claimed to be more expressive than four depthwise convolutions, with the proof relegated to the supplement.
  • IDA-SFT block
    purpose: Combines IDA with Spatial Feature Transform to synthesize both homogeneous and inhomogeneous degradations.
    Core building block of the synthesis network (Eq. 4); validated only through the paper's ablations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards a Universal Image Degradation Model via Content-Degradation Disentanglement." pith.science (2026). https://pith.science/paper/4AA5H5IM

@misc{pith2026250512860,
  author       = {Pith},
  title        = {Pith review of: Towards a Universal Image Degradation Model via Content-Degradation Disentanglement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4AA5H5IM}},
  note         = {Machine review of arXiv:2505.12860}
}
read the original abstract

Image degradation synthesis is highly desirable in a wide variety of applications ranging from image restoration to simulating artistic effects. Existing models are designed to generate one specific or a narrow set of degradations, which often require user-provided degradation parameters. As a result, they lack the generalizability to synthesize degradations beyond their initial design or adapt to other applications. Here we propose the first universal degradation model that can synthesize a broad spectrum of complex and realistic degradations containing both homogeneous (global) and inhomogeneous (spatially varying) components. Our model automatically extracts and disentangles homogeneous and inhomogeneous degradation features, which are later used for degradation synthesis without user intervention. A disentangle-by-compression method is proposed to separate degradation information from images. Two novel modules for extracting and incorporating inhomogeneous degradations are created to model inhomogeneous components in complex degradations. We demonstrate the model's accuracy and adaptability in film-grain simulation and blind image restoration tasks. The demo video, code, and dataset of this project will be released at github.com/yangwenbo99/content-degradation-disentanglement.

Figures

Figures reproduced from arXiv: 2505.12860 by the authors.

Figure 1
Figure 1. Demonstration of the applicability of our universal im [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Visualization of our models’ degradation disentanglement and transfer ability. Our model can extract homogeneous and inhomo [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Visual results for StyTr2 [16] retrained on the distortion dataset. It fails at separating distortion from content, hence, alters skin tone and texture while not transferring distortions (more ex￾amples in Supp. Sec. 3.5). Input distorted image Degradation encoding networks HDEN IDEN Clean image Random state Redistorted image Degradation synthesis network [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The overall architecture of our degradation encoding and [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]
Figure 5
Figure 5. Figure 5: The architecture of the two Degradation Encoding Net [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: The degradation synthesis network. cannot be reliably estimated, due to its high dimensional￾ity). We shall justify the entropy regularization loss’s effects in remainder of this section. Since el and eg are similar in nature, we shall drop the subscript until the disc…
Figure 7
Figure 7. Figure 7: Synthetic distortion transfer. Each grid is a synthesized [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Realistic distortion transfer on the raindrop dataset [ [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]
Figure 10
Figure 10. Figure 10: The result of film grain synthesis. Accur. ↓ Fidelity ↓ Realism ↓ w/o w/ w/o w/ w/o w/ Ours Ours Ours Ours Ours Ours 2 degradations NA .680 .513 .424 .334 228.8 28.1 AP .506 .463 .232 .219 45.1 16.9 UA .519 .512 .274 .275 47.8 30.0 NP .713 .485 .187 .081 221.9 20.6 UN…
Figure 11
Figure 11. Figure 11: Image restoration results for a single/complex com [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]
Figure 12
Figure 12. Figure 12: Examples of degradation transfer from mixed sources. [PITH_FULL_IMAGE:figures/full_fig_p008_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 45 canonical work pages

  1. [12]

    Bampis, Zhi Li, and Alan C

    Li-Heng Chen, Christos G. Bampis, Zhi Li, and Alan C. Bovik. Learning to distort images using generative adver- sarial networks.IEEE Signal Processing Letters, 27:2144– 2148, 2020. 2, 3, 6

    work page 2020

  2. [3]

    Style-based film grain analysis and synthesis

    Zoubida Ameur, Claire-H ´el`ene Demarty, Olivier Le Meur, Daniel M ´enard, and Edouard Franc ¸ois. Style-based film grain analysis and synthesis. InProceedings of the 14th ACM Multimedia Systems Conference, page 229–238, New York, NY , USA, 2023. Association for Computing Machinery. 7

    work page 2023

  3. [30]

    Robust un- supervised stylegan image restoration

    Yohan Poirier-Ginter and Jean-Franc ¸ois Lalonde. Robust un- supervised stylegan image restoration. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 22292–22301, 2023. 1, 7, 8

    work page 2023

  4. [1]

    Brubaker, and Michael S

    Abdelrahman Abdelhamed, Marcus A. Brubaker, and Michael S. Brown. Noise flow: Noise modeling with con- ditional normalizing flows. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2019. 1, 3

    work page 2019

  5. [2]

    Style-based film grain analysis and synthesis

    Zoubida Ameur, Claire-H ´el`ene Demarty, Olivier Le Meur, Daniel M ´enard, and Edouard Franc ¸ois. Style-based film grain analysis and synthesis. InProceedings of the 14th Con- ference on ACM Multimedia Systems, page 229–238, New York, NY , USA, 2023. Association for Computing Machin- ery. 3

    work page 2023

  6. [4]

    Deep-based film grain removal and synthesis.IEEE Transactions on Image Processing, 32:5046–5059, 2023

    Zoubida Ameur, Wassim Hamidouche, Edouard Franc ¸ois, Miloˇs Radosavljevi´c, Daniel Menard, and Claire-H´el`ene De- marty. Deep-based film grain removal and synthesis.IEEE Transactions on Image Processing, 32:5046–5059, 2023. 3

    work page 2023

  7. [5]

    Ancuti, Cosmin Ancuti, Chris Hermans, and Philippe Bekaert

    Codruta O. Ancuti, Cosmin Ancuti, Chris Hermans, and Philippe Bekaert. A fast semi-inverse approach to detect and remove the haze from a single image. InComputer Vision – ACCV 2010, pages 501–514, Berlin, Heidelberg, 2011. Springer Berlin Heidelberg. 3

    work page 2010

  8. [6]

    End-to-end optimized image compression.arXiv preprint arXiv:1611.01704, 2016

    Johannes Ball ´e, Valero Laparra, and Eero P Simoncelli. End-to-end optimized image compression.arXiv preprint arXiv:1611.01704, 2016. 4

    arXiv 2016

Show all 50 references
  1. [7]

    Variational image compres- sion with a scale hyperprior

    Johannes Ball ´e, David Minnen, Saurabh Singh, Sung Jin Hwang, and Nick Johnston. Variational image compres- sion with a scale hyperprior. InInternational Conference on Learning Representations, 2018. 4

    2018

  2. [8]

    Dehazenet: An end-to-end system for single image haze removal.IEEE transactions on image process- ing, 25(11):5187–5198, 2016

    Bolun Cai, Xiangmin Xu, Kui Jia, Chunmei Qing, and Dacheng Tao. Dehazenet: An end-to-end system for single image haze removal.IEEE transactions on image process- ing, 25(11):5187–5198, 2016. 1, 3

    2016

  3. [9]

    Im- age neural style transfer: A review.Computers and Electri- cal Engineering, 108:108723, 2023

    Qiang Cai, Mengxu Ma, Chen Wang, and Haisheng Li. Im- age neural style transfer: A review.Computers and Electri- cal Engineering, 108:108723, 2023. 3

    2023

  4. [10]

    Any-resolution training for high- resolution image synthesis

    Lucy Chai, Micha ¨el Gharbi, Eli Shechtman, Phillip Isola, and Richard Zhang. Any-resolution training for high- resolution image synthesis. InComputer Vision – ECCV 2022, pages 170–188, Cham, 2022. Springer Nature Switzer- land. 7

    2022

  5. [11]

    Learn- ing camera-aware noise models

    Ke-Chi Chang, Ren Wang, Hung-Jin Lin, Yu-Lun Liu, Chia- Ping Chen, Yu-Lin Chang, and Hwann-Tzong Chen. Learn- ing camera-aware noise models. InComputer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXIV, page 343–358, Berlin, ...

    2020

  6. [13]

    Diffusion pos- terior sampling for general noisy inverse problems

    Hyungjin Chung, Jeongsol Kim, Michael Thompson Mc- cann, Marc Louis Klasky, and Jong Chul Ye. Diffusion pos- terior sampling for general noisy inverse problems. InThe Eleventh International Conference on Learning Representa- tions, 2023. 7, 8

    2023

  7. [14]

    Stylegan-induced data-driven regularization for inverse problems

    Arthur Conmy, Subhadip Mukherjee, and Carola-Bibiane Sch¨onlieb. Stylegan-induced data-driven regularization for inverse problems. InICASSP 2022-2022 IEEE Interna- tional Conference on Acoustics, Speech and Signal Process- ing (ICASSP), pages 3788–3792. IEEE, 2022. 7

    2022

  8. [15]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255, 2009. 7

    2009

  9. [16]

    Stytr2: Im- age style transfer with transformers

    Yingying Deng, Fan Tang, Weiming Dong, Chongyang Ma, Xingjia Pan, Lei Wang, and Changsheng Xu. Stytr2: Im- age style transfer with transformers. InIEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 3

    2022

  10. [17]

    Simoncelli

    Keyan Ding, Kede Ma, Shiqi Wang, and Eero P. Simoncelli. Image quality assessment: Unifying structure and texture similarity.IEEE Transactions on Pattern Analysis and Ma- chine Intelligence, 44(5):2567–2581, 2022. 5

    2022

  11. [18]

    Practical poissonian-gaussian noise mod- eling and fitting for single-image raw-data.IEEE Transac- tions on Image Processing, 17(10):1737–1754, 2008

    Alessandro Foi, Mejdi Trimeche, Vladimir Katkovnik, and Karen Egiazarian. Practical poissonian-gaussian noise mod- eling and fitting for single-image raw-data.IEEE Transac- tions on Image Processing, 17(10):1737–1754, 2008. 3

    2008

  12. [19]

    Fre- quency separation for real-world super-resolution

    Manuel Fritsche, Shuhang Gu, and Radu Timofte. Fre- quency separation for real-world super-resolution. In2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), pages 3599–3608, 2019. 5

    2019

  13. [20]

    Commons:quality images - Wikimedia Commons.https://commons.wikimedia.org/ wiki/Commons:Quality_images, n.d

    Gnangarra, et al. Commons:quality images - Wikimedia Commons.https://commons.wikimedia.org/ wiki/Commons:Quality_images, n.d. 6

  14. [21]

    Single image haze removal using dark channel prior

    Kaiming He, Jian Sun, and Xiaoou Tang. Single image haze removal using dark channel prior. In2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 1956– 1963, 2009. 3

    1956

  15. [22]

    Real-world super-resolution via kernel estimation and noise injection

    Xiaozhong Ji, Yun Cao, Ying Tai, Chengjie Wang, Jilin Li, and Feiyue Huang. Real-world super-resolution via kernel estimation and noise injection. InThe IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition (CVPR) Workshops, 2020. 1, 3

    2020

  16. [23]

    Neural style transfer: A review

    Yongcheng Jing, Yezhou Yang, Zunlei Feng, Jingwen Ye, Yizhou Yu, and Mingli Song. Neural style transfer: A review. IEEE transactions on visualization and computer graphics, 26(11):3365–3385, 2019. 3

    2019

  17. [24]

    Noisetransfer: Image noise generation with contrastive embeddings

    Seunghwan Lee and Tae Hyun Kim. Noisetransfer: Image noise generation with contrastive embeddings. InProceed- ings of the Asian Conference on Computer Vision (ACCV), pages 3569–3585, 2022. 4

    2022

  18. [25]

    AOD-Net: All-in-one dehazing network

    Boyi Li, Xiulian Peng, Zhangyang Wang, Jizheng Xu, and Dan Feng. AOD-Net: All-in-one dehazing network. In2017 IEEE International Conference on Computer Vision (ICCV), pages 4780–4788, 2017. 3

    2017

  19. [26]

    Xuan Luo, Xuaner Cecilia Zhang, Paul Yoo, Ricardo Martin- Brualla, Jason Lawrence, and Steven M. Seitz. Time-travel rephotography.ACM Trans. Graph., 40(6), 2021. 3

    2021

  20. [27]

    Poisson-gaussian denoising using the exact unbiased inverse of the gener- alized anscombe transformation

    Markku M ¨akitalo and Alessandro Foi. Poisson-gaussian denoising using the exact unbiased inverse of the gener- alized anscombe transformation. In2012 IEEE Interna- tional Conference on Acoustics, Speech and Signal Process- ing (ICASSP), pages 1081–1084, 2012. 3

    2012

  21. [28]

    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. In2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 257–265, 2017. 5, 6

    2017

  22. [29]

    Film grain synthesis for av1 video codec

    Andrey Norkin and Neil Birkbeck. Film grain synthesis for av1 video codec. In2018 Data Compression Conference, pages 3–12, 2018. 1, 7

    2018

  23. [31]

    Robby T. Tan. Visibility in bad weather from a single image. In2008 IEEE Conference on Computer Vision and Pattern Recognition, pages 1–8, 2008. 3

    2008

  24. [32]

    From rain generation to rain re- moval

    Hong Wang, Zongsheng Yue, Qi Xie, Qian Zhao, Yefeng Zheng, and Deyu Meng. From rain generation to rain re- moval. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14791–14801, 2021. 1, 3

    2021

  25. [33]

    Unsuper- vised degradation representation learning for blind super- resolution

    Longguang Wang, Yingqian Wang, Xiaoyu Dong, Qingyu Xu, Jungang Yang, Wei An, and Yulan Guo. Unsuper- vised degradation representation learning for blind super- resolution. In2021 IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), pages 10576–10585,

  26. [34]

    Tianyu Wang, Xin Yang, Ke Xu, Shaozhe Chen, Qiang Zhang, and Rynson W.H. Lau. Spatial attentive single- image deraining with a high quality real rain dataset. In 2019 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 12262–12271, 2019. 5, 6

    2019

  27. [35]

    Recovering realistic texture in image super-resolution by deep spatial feature transform

    Xintao Wang, Ke Yu, Chao Dong, and Chen Change Loy. Recovering realistic texture in image super-resolution by deep spatial feature transform. InThe IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018. 4

    2018

  28. [36]

    Real-esrgan: Training real-world blind super-resolution with pure synthetic data

    Xintao Wang, Liangbin Xie, Chao Dong, and Ying Shan. Real-esrgan: Training real-world blind super-resolution with pure synthetic data. In2021 IEEE/CVF International Con- ference on Computer Vision Workshops (ICCVW), pages 1905–1914, 2021. 2

    1905

  29. [37]

    Bovik, H.R

    Zhou Wang, A.C. Bovik, H.R. Sheikh, and E.P. Simoncelli. Image quality assessment: from error visibility to structural similarity.IEEE Transactions on Image Processing, 13(4): 600–612, 2004. 5

    2004

  30. [38]

    Blind quality assessment of multiply distorted images using deep neural networks

    Zhongling Wang, Shahrukh Athar, and Zhou Wang. Blind quality assessment of multiply distorted images using deep neural networks. InImage Analysis and Recognition, pages 89–101, Cham, 2019. Springer International Publishing. 6

    2019

  31. [39]

    Information theoretical analysis of multi- variate correlation.IBM J

    Satosi Watanabe. Information theoretical analysis of multi- variate correlation.IBM J. Res. Dev., 4(1):66–82, 1960. 5

    1960

  32. [40]

    Physics-based noise modeling for extreme low-light photog- raphy.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):8520–8537, 2022

    Kaixuan Wei, Ying Fu, Yinqiang Zheng, and Jiaolong Yang. Physics-based noise modeling for extreme low-light photog- raphy.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):8520–8537, 2022. 3

    2022

  33. [41]

    J.C.K. Yan, P. Campisi, and D. Hatzinakos. Film grain noise removal and generation for color images. InPro- ceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP ’98 (Cat. No.98CH36181), pages 2957–2960 vol.5, 1998. 1, 3

    1998

  34. [42]

    Dual adversarial network: Toward real-world noise removal and noise generation

    Zongsheng Yue, Qian Zhao, Lei Zhang, and Deyu Meng. Dual adversarial network: Toward real-world noise removal and noise generation. InComputer Vision – ECCV 2020, pages 41–58, Cham, 2020. Springer International Publish- ing. 1, 3

    2020

  35. [43]

    Zeiler, Dilip Krishnan, Graham W

    Matthew D. Zeiler, Dilip Krishnan, Graham W. Taylor, and Rob Fergus. Deconvolutional networks. In2010 IEEE Com- puter Society Conference on Computer Vision and Pattern Recognition, pages 2528–2535, 2010. 4

    2010

  36. [44]

    Designing a practical degradation model for deep blind image super-resolution

    Kai Zhang, Jingyun Liang, Luc Van Gool, and Radu Timo- fte. Designing a practical degradation model for deep blind image super-resolution. InProceedings of the IEEE/CVF In- ternational Conference on Computer Vision (ICCV), pages 4791–4800, 2021. 2

    2021

  37. [45]

    Efros, Eli Shecht- man, and Oliver Wang

    Richard Zhang, Phillip Isola, Alexei A. Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. InProceedings of the IEEE Conference on Computer Vision and Pattern Recogni- tion (CVPR), 2018. 7

    2018

  38. [46]

    Rethinking noise synthesis and modeling in raw denois- ing

    Yi Zhang, Hongwei Qin, Xiaogang Wang, and Hongsheng Li. Rethinking noise synthesis and modeling in raw denois- ing. InProceedings of the IEEE/CVF International Confer- ence on Computer Vision (ICCV), pages 4593–4601, 2021. 1, 3

    2021

  39. [47]

    Joint learning con- tent and degradation aware feature for blind super-resolution

    Yifeng Zhou, Chuming Lin, Donghao Luo, Yong Liu, Ying Tai, Chengjie Wang, and Mingang Chen. Joint learning con- tent and degradation aware feature for blind super-resolution. InProceedings of the 30th ACM International Conference on Multimedia, page 2606–2616, New York, NY , U...

    2022

  40. [48]

    From noise modeling to blind image denoising

    Fengyuan Zhu, Guangyong Chen, and Pheng Ann Heng. From noise modeling to blind image denoising. In2016 IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), pages 420–429, 2016. 1, 3

    2016

  41. [49]

    A fast single image haze removal algorithm using color attenuation prior

    Qingsong Zhu, Jiaming Mai, and Ling Shao. A fast single image haze removal algorithm using color attenuation prior. IEEE Transactions on Image Processing, 24(11):3522–3533,

  42. [50]

    Estimating fine-grained noise model via contrastive learning

    Yunhao Zou and Ying Fu. Estimating fine-grained noise model via contrastive learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 12682–12691, 2022. 1, 3

    2022

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.