Pith. sign in

REVIEW 4 major objections 6 minor 67 references

CDI: Blind Image Restoration Fidelity Evaluation based on Consistency with Degraded Image

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A restored image's fidelity should be judged by how well it re-explains the degraded photo, not by similarity to a pristine reference; a wavelet consistency score does this without knowing the degradation.

desk verdict A promising but self-referential metric: the paper defines fidelity as consistency with the degraded image and then validates on a dataset that asks raters for exactly that judgment. read the letter →

arxiv 2501.14264 v2 pith:B7UXDYOK submitted 2025-01-24 eess.IV cs.CV

Xiaojun Tang , Jingru Wang , Guangwei Huang , Guannan Chen , Rui Zheng , Lian Huai , Yuyu Liu , Xingqun Jiang This is my paper · ORCID
classification eess.IVcs.CV MSC 68U1094A08
keywords blindimagerestorationqualityassessmentfidelityevaluationconsistencywithdegradedwaveletdomainreference-agnosticmetricdegraded-imagecomparisondatasetgenerative
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

Blind image restoration is ill-posed: from one degraded photo, many visually different restorations are all valid, so the pristine reference is only one solution and cannot be the ground truth for fidelity. The paper argues that fidelity should instead be measured as consistency with the degraded image (CDI), since degradation destroys high-frequency information irreversibly and only the low-frequency content can be checked. It introduces a wavelet-domain Reference Guided CDI (RGCDI) that removes noise from the degraded image, converts noise into an information-equivalent attenuation, and adaptively attenuates the restoration before computing PSNR. A reference-agnostic version (RACDI) learns the same attenuation directly from the degraded image, removing the need for a reference in real-world use. On a new subjective dataset (DISDCD), the CDI scores agree with human fidelity judgments far better than PSNR, SSIM, LPIPS, and DISTS.

What carries the argument

The load-bearing mechanism is the wavelet-domain consistency transform $F(x,y)=\mu_A x/\sqrt{1+\sigma_D^2/\sigma_H^2}$, built on a two-step degradation model $y=\mu_A x+n_D$ for wavelet coefficients and an HVS mutual-information model. Reference-guided noise removal splits the degraded image into an attenuated signal and independent noise, the noise-equivalent attenuation step turns noise-induced information loss into an equivalent extra attenuation, and adaptive attenuation picks the least-squares coefficient $\mu_M$ that best fits the restoration to the attenuated reference before computing PSNR. RGCDI's idempotency and cascade-degradation properties make the transform stable under repeated and combined degradations. The reference-agnostic version replaces the reference-guided noise removal with a denoising backbone trained to predict $F(x,y)$ from the degraded image alone, so consistency becomes computable without a reference.

What would settle it

A decisive test would be a preference study where raters choose which restoration is more faithful to the original scene, with the pristine original shown as reference, among restorations that have identical low-frequency consistency with the degraded image but different high-frequency realism; if raters consistently prefer one, CDI misses part of fidelity. A simpler check is to compute CDI on a dataset whose ground truth was collected against the pristine original rather than by re-degrading restorations.

Watch

Extended reading notes

Core claim

The paper's central claim is that full-reference fidelity evaluation is the wrong yardstick for blind image restoration. Because the original image is lost during degradation, any restoration that re-degrades to the observed degraded image is a legitimate solution, and high-frequency details generated by restoration are unverifiable in principle. The paper therefore defines fidelity as the consistency between the restored image and the degraded image in low-frequency content, and constructs RGCDI to compute that consistency without knowing the degradation parameters: it decomposes the degraded wavelet coefficients into attenuation plus noise, converts the noise into equivalent attenuation through an HVS mutual-information model, adaptively attenuates the restoration to best match the attenuated reference, and reports PSNR in the image domain. RACDI replaces the reference-guided part with a trained attenuation network, so the same consistency score is available without any reference image. The paper's experiments on DISDCD show these scores track human fidelity choices substantially better than PSNR, SSIM, LPIPS, and DISTS, and also better than running those full-reference metrics on the re-degraded image.

Load-bearing premise

The load-bearing premise is that DISDCD's subjective choices—'which re-degraded restoration looks closer to the degraded image'—are a valid operationalization of fidelity; if human fidelity judgment includes anything beyond low-frequency closeness to the degraded image, the experimental support weakens.

Editorial extensions

If this is right

  • RGCDI and RACDI reach higher 2AFC agreement with human fidelity judgments on DISDCD than PSNR, SSIM, LPIPS, DISTS, and FR-IQA computed on re-degraded images.
  • Full-reference scores demonstrably reward blurring a restoration, so evaluation pipelines built on them will push blind restoration methods toward smoother, less faithful outputs.
  • RACDI's mean error relative to RGCDI stays below roughly 2.7 dB across deblurring, super-resolution, denoising, and blind restoration tasks, making reference-free fidelity evaluation practical.
  • Because CDI deliberately ignores generated high-frequency content, the paper positions it as orthogonal to no-reference quality assessment; a complete BIR evaluation would supplement a CDI fidelity score with an NR-IQA perceptual quality score.

Reading between the lines

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

  • Editorial inference: DISDCD asks raters which re-degraded restoration is closer to the degraded image, which is exactly what CDI measures; a fidelity benchmark where raters compare against the pristine original could give reference-based metrics more credit than this protocol does.
  • Editorial inference: The paper does not test combining CDI with NR-IQA, though its orthogonality claim suggests a two-axis evaluation (fidelity versus perceptual quality) rather than a single score.
  • Editorial inference: The proven inequality RGCDI PSNR >= PSNR means CDI values are on a different scale from classical PSNR; adopters should compare CDI differences within one evaluation rather than against published PSNR tables.
  • Editorial inference: If CDI is used as a training loss, it will reward restorations that re-degrade to the input and ignore the realism of invented detail, an interaction with the perception-distortion tradeoff the paper does not discuss.
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 / 6 minor

Summary. The paper proposes a new fidelity metric for blind image restoration (BIR) that compares a restored image with the degraded image rather than with the reference image. The method, called Consistency with Degraded Image (CDI), is implemented in two variants: RGCDI, a wavelet-domain algorithm that removes noise from the degraded image with reference guidance and applies information-equivalent attenuation, and RACDI, a reference-agnostic variant that approximates RGCDI using a trained Attenuation Net. The authors construct a new subjective dataset, DISDCD, based on pairwise (2AFC) judgments of which re-degraded restored image is closer to the degraded image, and report that RGCDI and RACDI achieve higher agreement with these judgments than existing FR-IQA and NR-IQA methods.

Significance. The core idea of anchoring fidelity evaluation on the degraded image is practically appealing for BIR, where a reference is unavailable. The information-theoretic derivation in Sec. 4, in particular the equivalence between additive noise and equivalent signal attenuation (Eq. 4c = 4d), is a clean theoretical contribution under the stated Gaussian assumptions, and the proposed algorithm has attractive invariance properties if they are correctly proven. The DISDCD dataset is a potentially useful resource for the community, although its design currently conflates the metric's target with the ground-truth protocol. If the authors can validate CDI against an independent measure of human fidelity judgment, the approach would be a valuable complement to existing IQA methods for BIR.

major comments (4)
  1. [Sec. 7 and Sec. 8.1, Fig. 11] The DISDCD subjective task asks raters to choose which of two re-degraded restored images is closer to the degraded image. This is precisely the quantity that RGCDI computes: RGCDI adaptively attenuates the restored image (Eq. 5) and compares it with the noise-removed, equivalently attenuated degraded image (Sec. 5). Consequently, the high 2AFC agreement in Fig. 11 largely verifies that RGCDI approximates its own target, not that the CDI construct is the correct operationalization of BIR fidelity. The paper must validate CDI against an independent subjective protocol, for example a preference task where raters rank restored images by fidelity to a reference (when available) or by overall perceptual quality, and demonstrate that CDI-predicted rankings correlate with those judgments.
  2. [Sec. 5.3, Fig. 12 and Sec. 6, Table 1] The hyperparameter λ is tuned by maximizing 2AFC agreement on DISDCD (Fig. 12), and the recommended value of 0.3 is then used for the reported comparisons. Since DISDCD is the same benchmark on which the metric is evaluated, this tuning does not provide independent evidence of generalization. Similarly, RACDI is trained with RGCDI-generated pseudo-labels (Sec. 6) and evaluated only against RGCDI (Table 1), so the reported mean errors measure approximation of RGCDI rather than alignment with human fidelity judgments. The claim of 'significantly superior' would be supported by evaluation on at least one external dataset or a held-out portion of DISDCD using a distinct subjective protocol.
  3. [Sec. 8.1, Fig. 11] Fig. 11 reports point estimates of 2AFC accuracy without error bars, confidence intervals, or any significance test. The number of raters in DISDCD is not stated, and no inter-rater agreement statistics are provided. The observation that RGCDI scores 'even exceed human scores' is not discussed; a deterministic metric will often exceed the human consistency measure of Eq. 9, which is the expected agreement between two raters, and this should be interpreted carefully. A proper statistical comparison with confidence intervals and a test of differences between methods is needed before the word 'significantly' is used.
  4. [Sec. 1.2 and Sec. 9] The paper explicitly redefines BIR fidelity as low-frequency consistency with the degraded image and states that CDI does not evaluate high-frequency signals, which are left to NR-IQA. This is a legitimate design choice, but it changes the meaning of 'fidelity' relative to the standard FR-IQA notion. The conclusion that CDI is 'markedly superior' to FR-IQA methods is then partly a consequence of measuring a different construct. The paper should prominently and consistently frame CDI as a new, complementary evaluation axis rather than as a general replacement for fidelity assessment, and the experimental comparisons should be interpreted in that light.
minor comments (6)
  1. [Abstract, Sec. 4.2, Sec. 6, Fig. 4] There are several typos and inconsistent terms: 'In stead' in the abstract, 'Gussian' in the Fig. 4 caption, 'HSV' should be 'HVS' in Sec. 4.2, and 'Noise Equivalent Attention' should be 'Noise Equivalent Attenuation' in Sec. 6.
  2. [Sec. 7] The DISDCD dataset description lacks important experimental details: the number of raters, rater screening and training, the number of trials per rater, and the exact meaning of '5 pairwise judgments for each sample' need to be specified so that the reliability of the subjective ground truth can be assessed.
  3. [Sec. 8.1] The comparison with NR-IQA methods (NIQE, MANIQA) is not a like-for-like comparison because these methods measure perceptual naturalness, not fidelity. The statement that they 'cannot evaluate fidelity' is true, but the figure and text should acknowledge that the comparison spans different constructs.
  4. [Supplementary Material, Sec. 12, Eq. 17] The proof of idempotency for the case F(F(x,y), y) contains an algebraic error: when computing μ'_A F(x,y) = COV(y, F(x,y)) / COV(F(x,y), F(x,y)), the factor α = μ_A / sqrt(1 + σ_D²/σ_H²) is omitted, leading to the incorrect intermediate assertion μ'_A F(x,y) = μ_A x. The final equality F(F(x,y), y) = F(x,y) holds only after the correction, so the derivation as written should be revised.
  5. [Abstract and Sec. 7] The paper states that the source code and DISDCD dataset will be publicly available 'shortly'. For reproducibility, please provide a URL or a clear availability statement, and include the dataset creation details (e.g., degradation parameters, seed handling, and image sources) in the supplementary material.
  6. [Table 1 and Tables 8-10] The reported mean errors between RACDI and RGCDI lack standard deviations and sample sizes. Adding these statistics would help the reader judge the consistency of the approximation across different tasks and images.

Circularity Check

2 steps flagged · score 6.0 of 10

DISDCD's ground truth asks raters for the same quantity RGCDI computes, so the headline comparison is largely self-fulfilling; RACDI adds no independent validation because it is trained on RGCDI-generated targets.

  1. self definitional [Sec. 7 (DISDCD annotation task) vs. Sec. 5.1 (RGCDI definition, Eq. 5)]
    "To better observe the differences between the degraded images, we apply the same degradation operation on the two restored images, producing the degraded restored images. Then a rater is asked to determine which is closer to the degraded image, and the response is recorded. ... Therefore, CDI should disregard high-frequency information and focus on assessing the consistency of low-frequency information."

    RGCDI's score is PSNR(IDWT(µM x), IDWT(µAx/sqrt(1+σD^2/σH^2))) (Eq.5): it adaptively attenuates the restored image and compares it with an attenuated, noise-removed version of the degraded image. DISDCD asks raters to re-degrade two restored images with the same degradation and choose which is closer to the degraded image. The paper further defines fidelity as low-frequency consistency with the degraded image (Sec.1.2). So the subjective labels operationalize the same construct RGCDI computes; high 2AFC agreement largely verifies RGCDI matches its own target, not that this target is independently the correct meaning of BIR fidelity. The comparison is partly self-fulfilling, though the implementation is nontrivial.

  2. fitted input called prediction [Sec. 6 and Sec. 8.3, Table 1]
    "paired ( y, µAx/p 1 + σ2 D/σ2 H ) training images of various types of degradation can be generated through the RGCDI system. After end-to-end training, the Attenuation Net is able to predictµAx/p 1 + σ2 D/σ2 H from y. ... the number on the left represents RACDI PSNR and the number on the right represents the mean error of RACDIPSNR relative to RGCDI PSNR."

    RACDI is trained end-to-end on RGCDI-generated pairs (y, µAx/sqrt(1+σD^2/σH^2)), and Table 1 reports 'the mean error of RACDIPSNR relative to RGCDI PSNR'. Small error therefore measures how well the attenuation net regresses its own training target; it is an approximation check, not an independent fidelity validation. RACDI's DISDCD score (Fig.11) inherits the same construct-alignment between RGCDI and DISDCD. Presenting this as 'enabling BIR fidelity evaluation without reference images' is an engineering claim, not additional empirical support for CDI.

full rationale

The paper's analytic core is largely self-contained: the HVS information model (Eqs.4a-4d) is a straightforward derivation, the noise-equivalent attenuation identity is algebraic, and the idempotency/cascade/PSNR-inequality properties follow from the stated definitions in the supplementary material. There is no load-bearing self-citation chain. The circularity is in the external-validity argument: DISDCD labels are constructed as the very consistency judgment CDI computes, and RACDI is trained and evaluated against RGCDI. Because the paper's headline claim of 'significantly superior' fidelity evaluation rests on DISDCD results, the empirical support is partly self-fulfilling. If one accepts the paper's own definition of fidelity as low-frequency consistency with the degraded image, RGCDI is a reasonable implementation; but the dataset cannot independently establish that definition against alternative fidelity notions such as high-frequency plausibility, semantic correctness, or naturalness.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a chain of modeling assumptions: a two-step degradation model with independent additive noise, a Gaussian scale mixture prior for wavelet coefficients, an additive Gaussian noise model for HVS, and a subjective dataset whose definition of fidelity matches the metric. The only manually tuned scalar is lambda; the Attenuation Net adds millions of learned parameters trained on RGCDI-generated targets.

free parameters (3)
  • lambda (HVS noise scale) = 0.3 (recommended)
    Only manually adjusted hyperparameter; sets sigma_H squared proportional to attenuated reference variance, Eq.8; recommended value selected on DISDCD (Fig.12).
  • Attenuation Net weights = trained network weights (SCUNet backbone)
    End-to-end training on RGCDI-generated pairs (Sec.6) makes RACDI's predictions dependent on millions of fitted parameters.
  • Per-image attenuation coefficient mu_A and noise variance sigma_D squared = estimated per image from Eq.3
    Estimated from covariance between reference and degraded wavelet coefficients; not free hyperparameters, but data-fitted quantities used in the score formula.
assumptions (6)
  • domain assumption Image degradation can be modeled as y = mu_A x + n_D with noise n_D independent of the signal x (Eq.1, Sec.4.1).
    Used to split degradation into attenuation and noise, and to justify mu_A and sigma_D squared estimates in Eq.3.
  • domain assumption Wavelet coefficients of natural images follow a Gaussian Scale Mixture (Eq.2, Sec.4.1).
    Borrowed from [35] to make the Gaussian mutual information calculations in Eq.4 valid.
  • domain assumption The HVS can be modeled as an additive Gaussian noise channel with variance sigma_H squared (Eq.4, Sec.4.2).
    Taken from [32]; underpins the noise-equivalent attenuation used to handle noisy degradations.
  • ad hoc to paper Information-equivalent attenuation preserves perceived fidelity; replacing noise with attenuation that gives the same mutual information is valid (Eq.4c=4d).
    This equivalence is derived for Gaussian variables, but the paper extends it to perceptual fidelity without human validation.
  • domain assumption Re-degrading restored images with the same degradation, with fixed random seed, gives a fair subjective comparison in DISDCD (Sec.7).
    Used to construct the 2AFC task; assumes the degradation process is repeatable and that re-degraded images preserve the fidelity signal.
  • standard math Orthogonal wavelet transform preserves L2 distance (Supplementary, Eq.22-24).
    Used to justify comparing coefficients instead of pixel images in RGCDI PSNR.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CDI: Blind Image Restoration Fidelity Evaluation based on Consistency with Degraded Image." pith.science (2026). https://pith.science/paper/B7UXDYOK

@misc{pith2026250114264,
  author       = {Pith},
  title        = {Pith review of: CDI: Blind Image Restoration Fidelity Evaluation based on Consistency with Degraded Image},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B7UXDYOK}},
  note         = {Machine review of arXiv:2501.14264}
}
read the original abstract

Recent advancements in Blind Image Restoration (BIR) methods, based on Generative Adversarial Networks and Diffusion Models, have significantly improved visual quality. However, they present significant challenges for Image Quality Assessment (IQA), as the existing Full-Reference IQA methods often rate images with high perceptual quality poorly. In this paper, we reassess the Solution Non-Uniqueness and Degradation Indeterminacy issues of BIR, and propose constructing a specific BIR IQA system. In stead of directly comparing a restored image with a reference image, the BIR IQA evaluates fidelity by calculating the Consistency with Degraded Image (CDI). Specifically, we propose a wavelet domain Reference Guided CDI algorithm, which can acquire the consistency with a degraded image for various types without requiring knowledge of degradation parameters. The supported degradation types include down sampling, blur, noise, JPEG and complex combined degradations etc. In addition, we propose a Reference Agnostic CDI, enabling BIR fidelity evaluation without reference images. Finally, in order to validate the rationality of CDI, we create a new Degraded Images Switch Display Comparison Dataset (DISDCD) for subjective evaluation of BIR fidelity. Experiments conducted on DISDCD verify that CDI is markedly superior to common Full Reference IQA methods for BIR fidelity evaluation. The source code and the DISDCD dataset will be publicly available shortly.

Figures

Figures reproduced from arXiv: 2501.14264 by the authors.

Figure 1
Figure 1. (a) Classic IQA includes Full Reference IQA and No [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. BIR resolution is non-unique. It is impossible to de [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 4
Figure 4. (a) BIR Gussian blur test system. (b, c) Appropri [PITH_FULL_IMAGE:figures/full_fig_p002_4.png] view at source ↗
Figures from the paper (7 more)
Figure 6
Figure 6. Figure 6: Two-Step Image Degradation Modeling According to Eq.1, we aim to decompose y into µAx and nD. Given reference x, the noise separation is straightfor￾ward, as COV (nD, x) = 0 (that is, x and nD are indepen￾dent in probability [32], see Eq.3). The standard deviation σn c…
Figure 5
Figure 5. Figure 5: FR-IQA on Degraded Image 4. Fidelity Subjective Evaluation Modeling To analyze the impact of noise on fidelity evaluation, we first use a two-step image degradation model to separate im￾age degradation into image attenuation and noise (Sec.4.1). We then examine the eff…
Figure 7
Figure 7. Figure 7: Example sample images of Wavelet Noise Splitting [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]
Figure 9
Figure 9. Figure 9: (a) Reference Guided CDI calculates the wavelet domain [PITH_FULL_IMAGE:figures/full_fig_p006_9.png]
Figure 10
Figure 10. Figure 10: DISDCD annotation interface BIR restoration algorithms include BSRGAN [52], Real￾ESRGAN [37], LDL [21], DASR [20], FeMaSR [4], LDM [31], StableSR [36], ResShift [47], PASD [44], DiffBIR [22] and SeeSR [41]. The conventional method of comparing left and right im￾ages m…
Figure 12
Figure 12. Figure 12: Hyper parameter (λ) tuning [PITH_FULL_IMAGE:figures/full_fig_p007_12.png]
Figure 13
Figure 13. Figure 13: Attenuation Net Backbones Comparison. 8.1. Comparison of IQA Methods on DISDCD This experiment compares the proposed RGCDI and RACDI with common FR-IQA methods (including PSNR, SSIM [39], LPIPS [58], DISTS [10]), FR-IQA on degraded images and human judgments. DEG P SN…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 60 canonical work pages

  1. [1]

    Recommendation itu-r bt.500-15, 2023. 7

    work page 2023

  2. [2]

    Ntire 2017 challenge on single image super-resolution: Dataset and study

    Eirikur Agustsson and Radu Timofte. Ntire 2017 challenge on single image super-resolution: Dataset and study. In2017 IEEE Conference on Computer Vision and Pattern Recogni- tion Workshops (CVPRW), pages 1122–1131, 2017. 2, 6

    work page 2017

  3. [3]

    The perception-distortion tradeoff

    Yochai Blau and Tomer Michaeli. The perception-distortion tradeoff. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6228–6237, 2018. 3

    work page 2018

  4. [4]

    Real-world blind super-resolution via feature matching with implicit high- resolution priors

    Chaofeng Chen, Xinyu Shi, Yipeng Qin, Xiaoming Li, Xi- aoguang Han, Tao Yang, and Shihui Guo. Real-world blind super-resolution via feature matching with implicit high- resolution priors. 2022. 7, 8

    work page 2022

  5. [5]

    Pre-trained image processing transformer

    Hanting Chen, Yunhe Wang, Tianyu Guo, Chang Xu, Yip- ing Deng, Zhenhua Liu, Siwei Ma, Chunjing Xu, Chao Xu, and Wen Gao. Pre-trained image processing transformer. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), pages 12299–12310,

    work page

  6. [6]

    Zhang, and Jian Sun

    Liangyu Chen, Xiaojie Chu, X. Zhang, and Jian Sun. Simple baselines for image restoration. In European Conference on Computer Vision, 2022. 8

    work page 2022

  7. [7]

    Exploring efficient asymmetric blind-spots for self- supervised denoising in real-world scenarios

    Shiyan Chen, Jiyuan Zhang, Zhaofei Yu, and Tiejun Huang. Exploring efficient asymmetric blind-spots for self- supervised denoising in real-world scenarios. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 2814–2823, 2024. 8

    work page 2024

  8. [8]

    Rethinking coarse-to-fine approach in sin- gle image deblurring

    Sung-Jin Cho, Seoyoun Ji, Jun-Pyo Hong, Seung-Won Jung, and Sung-Jea Ko. Rethinking coarse-to-fine approach in sin- gle image deblurring. 2021 IEEE/CVF International Confer- ence on Computer Vision (ICCV), pages 4621–4630, 2021. 8

    work page 2021

Show all 67 references
  1. [9]

    Second-order attention network for single image super-resolution

    Tao Dai, Jianrui Cai, Yongbing Zhang, Shu-Tao Xia, and Lei Zhang. Second-order attention network for single image super-resolution. In 2019 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 11057– 11066, 2019. 1, 8

    2019

  2. [10]

    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. 2, 3, 7

    2022

  3. [11]

    Davis, and Abhinav Shrivastava

    Max Ehrlich, Ser-Nam Lim, Larry S. Davis, and Abhinav Shrivastava. Quantization guided jpeg artifact correction. In European Conference on Computer Vision, 2020. 8

    2020

  4. [12]

    Jinjin Gu, Haoming Cai, Haoyu Chen, Xiaoxing Ye, Jimmy S. J. Ren, and Chao Dong. Pipal: a large-scale image qual- ity assessment dataset for perceptual image restoration. In European Conference on Computer Vision, 2020. 3, 6

    2020

  5. [13]

    Towards flexible blind jpeg artifacts removal

    Jiaxi Jiang, Kai Zhang, and Radu Timofte. Towards flexible blind jpeg artifacts removal. In 2021 IEEE/CVF Interna- tional Conference on Computer Vision (ICCV), pages 4977– 4986, 2021. 8

    2021

  6. [14]

    Beyond color difference: Residual in- terpolation for color image demosaicking

    Daisuke Kiku, Yusuke Monno, Masayuki Tanaka, and Masatoshi Okutomi. Beyond color difference: Residual in- terpolation for color image demosaicking. IEEE Transac- tions on Image Processing, 25(3):1288–1300, 2016. 1, 8

    2016

  7. [15]

    Larson and Damon M

    Eric C. Larson and Damon M. Chandler. Most apparent dis- tortion: full-reference image quality assessment and the role of strategy. J. Electronic Imaging, 19:011006, 2010. 3, 6

    2010

  8. [16]

    A simple baseline for video restoration with grouped spatial- temporal shift

    Dasong Li, Xiaoyu Shi, Yi Zhang, Ka Chun Cheung, Simon See, Xiaogang Wang, Hongwei Qin, and Hongsheng Li. A simple baseline for video restoration with grouped spatial- temporal shift. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 9822–9832,

    2023

  9. [17]

    Spatially adap- tive self-supervised learning for real-world image denoising

    Junyi Li, Zhilu Zhang, Xiaoyu Liu, Chaoyu Feng, Xiao- tao Wang, Lei Lei, and Wangmeng Zuo. Spatially adap- tive self-supervised learning for real-world image denoising. 2023 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 9914–9924, 2023. 8

    2023

  10. [18]

    Efficient and explicit modelling of image hierarchies for image restora- tion

    Yawei Li, Yuchen Fan, Xiaoyu Xiang, Denis Demandolx, Rakesh Ranjan, Radu Timofte, and Luc Gool. Efficient and explicit modelling of image hierarchies for image restora- tion. pages 18278–18289, 2023. 8

    2023

  11. [19]

    Swinir: Image restoration using swin transformer

    Jingyun Liang, Jiezhang Cao, Guolei Sun, Kai Zhang, Luc Van Gool, and Radu Timofte. Swinir: Image restoration using swin transformer. arXiv preprint arXiv:2108.10257 ,

    arXiv

  12. [20]

    Efficient and degradation-adaptive network for real-world image super- resolution

    Jie Liang, Hui Zeng, and Lei Zhang. Efficient and degradation-adaptive network for real-world image super- resolution. In European Conference on Computer Vision ,

  13. [21]

    Details or artifacts: A locally discriminative learning approach to realistic image super-resolution

    Jie Liang, Huiyu Zeng, and Lei Zhang. Details or artifacts: A locally discriminative learning approach to realistic image super-resolution. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 5647–5656,

    2022

  14. [22]

    Qiao, and Chao Dong

    Xin Yu Lin, Jingwen He, Zi-Yuan Chen, Zhaoyang Lyu, Ben Fei, Bo Dai, Wanli Ouyang, Y . Qiao, and Chao Dong. Diff- bir: Towards blind image restoration with generative diffu- sion prior. ArXiv, abs/2308.15070, 2023. 1, 2, 7, 8

    2023 arXiv

  15. [23]

    Invertible de- noising network: A light solution for real noise removal

    Yang Liu, Zhenyue Qin, Saeed Anwar, Pan Ji, Dongwoo Kim, Sabrina Caldwell, and Tom Gedeon. Invertible de- noising network: A light solution for real noise removal. In 2021 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 13360–13369, 2021. 8

    2021

  16. [24]

    No-reference image quality assessment in the spa- tial domain

    Anish Mittal, Anush Krishna Moorthy, and Alan Conrad Bovik. No-reference image quality assessment in the spa- tial domain. IEEE Transactions on Image Processing , 21 (12):4695–4708, 2012. 3

    2012

  17. [25]

    Single Image Super-Resolution via a Holistic Attention Network, pages 191–207

    Ben Niu, Weilei Wen, Wenqi Ren, Xiangde Zhang, Lianping Yang, Shuzhen Wang, Kaihao Zhang, Xiaochun Cao, and Haifeng Shen. Single Image Super-Resolution via a Holistic Attention Network, pages 191–207. 2020. 1, 8

    2020

  18. [26]

    Random sub-samples generation for self- supervised real image denoising

    Yizhong Pan, Xiao Liu, Xiangyu Liao, Yuanzhouhan Cao, and Chao Ren. Random sub-samples generation for self- supervised real image denoising. 2023 IEEE/CVF Inter- national Conference on Computer Vision (ICCV) , pages 12116–12125, 2023. 8

    2023

  19. [27]

    Multi-temporal recurrent neural networks for progres- sive non-uniform single image deblurring with incremental temporal training

    Dongwon Park, Dong un Kang, Jisoo Kim, and Se Young Chun. Multi-temporal recurrent neural networks for progres- sive non-uniform single image deblurring with incremental temporal training. In European Conference on Computer Vi- sion, 2019. 8

    2019

  20. [28]

    Nikolay Ponomarenko, Lina Jin, Oleg Ieremeiev, Vladimir Lukin, Karen Egiazarian, Jaakko Astola, Benoit V ozel, Kacem Chehdi, Marco Carli, Federica Battisti, and C.-C. Jay Kuo. Image database tid2013: Peculiarities, results and per- spectives. Signal Processing: Image Communica...

    2015

  21. [29]

    Ponomarenko, Vladimir V

    Nikolay N. Ponomarenko, Vladimir V . Lukin, and Alexan- der A. Zelensky. Tid2008 – a database for evaluation of full- reference visual quality assessment metrics. 2004. 3

    2004

  22. [30]

    Pieapp: Perceptual image-error assessment through pairwise preference

    Ekta Prashnani, Hong Cai, Yasamin Mostofi, and Pradeep Sen. Pieapp: Perceptual image-error assessment through pairwise preference. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 1808– 1817, 2018. 3, 6

    2018

  23. [31]

    Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer

    Robin Rombach, A. Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10674–10685, 2021. 7

    2022

  24. [32]

    Sheikh and A.C

    H.R. Sheikh and A.C. Bovik. Image information and visual quality. IEEE Transactions on Image Processing, 15(2):430– 444, 2006. 4, 6

    2006

  25. [33]

    Sheikh, M.F

    H.R. Sheikh, M.F. Sabir, and A.C. Bovik. A statistical eval- uation of recent full reference image quality assessment al- gorithms. IEEE Transactions on Image Processing, 15(11): 3440–3451, 2006. 3, 6

    2006

  26. [34]

    Rajiv Soundararajan and Alan C. Bovik. Rred indices: Re- duced reference entropic differencing for image quality as- sessment. IEEE Transactions on Image Processing , 21(2): 517–526, 2012. 3

    2012

  27. [35]

    Wainwright, Eero P

    Martin J. Wainwright, Eero P. Simoncelli, and Alan S. Will- sky. Random cascades on wavelet trees and their use in ana- lyzing and modeling natural images. Applied and Computa- tional Harmonic Analysis, 11(1):89–123, 2001. 4

    2001

  28. [36]

    Chan, and Chen Change Loy

    Jianyi Wang, Zongsheng Yue, Shangchen Zhou, Kelvin C.K. Chan, and Chen Change Loy. Exploiting diffusion prior for real-world image super-resolution. 2024. 1, 2, 7, 8

    2024

  29. [37]

    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. In 2021 IEEE/CVF International Con- ference on Computer Vision Workshops (ICCVW) , pages 1905–1914, 2021. 1, 2, 4, 6, 7, 8

    2021

  30. [38]

    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, abs/2212.00490, 2022. 4

    2022 arXiv

  31. [39]

    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. 2, 3, 7

    2004

  32. [40]

    Orientation selectivity based visual pattern for reduced-reference image quality assessment

    Jinjian Wu, Weisi Lin, Guangming Shi, Leida Li, and Yum- ing Fang. Orientation selectivity based visual pattern for reduced-reference image quality assessment. Information Sciences, 351:18–29, 2016. 3

    2016

  33. [41]

    Seesr: Towards semantics- aware real-world image super-resolution

    Rongyuan Wu, Tao Yang, Lingchen Sun, Zhengqiang Zhang, Shuai Li, and Lei Zhang. Seesr: Towards semantics- aware real-world image super-resolution. In2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 25456–25467, 2024. 1, 2, 7, 8

    2024

  34. [42]

    Cgnet: A light-weight context guided network for semantic segmentation

    Tianyi Wu, Sheng Tang, Rui Zhang, Juan Cao, and Yong- dong Zhang. Cgnet: A light-weight context guided network for semantic segmentation. IEEE Transactions on Image Processing, 30:1169–1179, 2021. 8

    2021

  35. [43]

    Maniqa: Multi-dimension at- tention network for no-reference image quality assessment

    Sidi Yang, Tianhe Wu, Shu Shi, Shan Gong, Ming Cao, Ji- ahao Wang, and Yujiu Yang. Maniqa: Multi-dimension at- tention network for no-reference image quality assessment. 2022 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition Workshops (CVPRW) , pages 1190–1199,

    2022

  36. [44]

    Pixel- aware stable diffusion for realistic image super-resolution and personalized stylization

    Tao Yang, Peiran Ren, Xuansong Xie, and Lei Zhang. Pixel- aware stable diffusion for realistic image super-resolution and personalized stylization. ArXiv, abs/2308.14469, 2023. 7

    2023 arXiv

  37. [45]

    Variational denoising network: Toward blind noise modeling and removal

    Zongsheng Yue, Hongwei Yong, Qian Zhao, Lei Zhang, and Deyu Meng. Variational denoising network: Toward blind noise modeling and removal. ArXiv, abs/1908.11314, 2019. 8

    1908 arXiv

  38. [46]

    Dual adversarial network: Toward real-world noise removal and noise generation, 2020

    Zongsheng Yue, Qian Zhao, Lei Zhang, and Deyu Meng. Dual adversarial network: Toward real-world noise removal and noise generation, 2020. 8

    2020

  39. [47]

    Resshift: Efficient diffusion model for image super- resolution by residual shifting.ArXiv, abs/2307.12348, 2023

    Zongsheng Yue, Jianyi Wang, and Chen Change Loy. Resshift: Efficient diffusion model for image super- resolution by residual shifting.ArXiv, abs/2307.12348, 2023. 7

    2023 arXiv

  40. [48]

    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 - 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022 , page...

    2022

  41. [49]

    Multi-stage progressive image restoration

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

    2021

  42. [50]

    Deep stacked hierarchical multi-patch network for im- age deblurring

    Hongguang Zhang, Yuchao Dai, Hongdong Li, and Piotr Ko- niusz. Deep stacked hierarchical multi-patch network for im- age deblurring. In 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 5971–5979,

    2019

  43. [51]

    Blur- aware spatio-temporal sparse transformer for video deblur- ring

    Huicong Zhang, Haozhe Xie, and Hongxun Yao. Blur- aware spatio-temporal sparse transformer for video deblur- ring. 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3616–3626, 2024. 1

    2024

  44. [52]

    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. In IEEE International Conference on Computer Vision, pages 4791–4800, 2021. 1, 2, 7, 8

    2021

  45. [53]

    Practical blind denoising via swin-conv-unet and data synthesis, 2022

    Kai Zhang, Yawei Li, Jingyun Liang, Jiezhang Cao, Yulun Zhang, Hao Tang, Radu Timofte, and Luc Gool. Practical blind denoising via swin-conv-unet and data synthesis, 2022. 6, 8

    2022

  46. [55]

    Plug-and-play image restora- tion with deep denoiser prior

    Kai Zhang, Yawei Li, Wangmeng Zuo, Lei Zhang, Luc Van Gool, and Radu Timofte. Plug-and-play image restora- tion with deep denoiser prior. IEEE Transactions on Pat- tern Analysis and Machine Intelligence , 44(10):6360–6376,

  47. [56]

    Lin Zhang, Lei Zhang, and Alan C. Bovik. A feature- enriched completely blind image quality evaluator. IEEE Transactions on Image Processing, 24(8):2579–2591, 2015. 3

    2015

  48. [57]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In CVPR, 2018. 3, 6

    2018

  49. [58]

    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. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 586–595, 2018. 2, 3, 7, 8

    2018

  50. [59]

    Image super-resolution using very deep residual channel attention networks

    Yulun Zhang, Kunpeng Li, Kai Li, Lichen Wang, Bineng Zhong, and Yun Fu. Image super-resolution using very deep residual channel attention networks. In Computer Vision – ECCV 2018: 15th European Conference, Munich, Ger- many, September 8–14, 2018, Proceedings, Part VII , page 2...

    2018

  51. [60]

    Vi- sual quality assessment for super-resolved images: Database and method

    Fei Zhou, Rongguo Yao, Bozhi Liu, and Guoping Qiu. Vi- sual quality assessment for super-resolved images: Database and method. IEEE Transactions on Image Processing, 28(7): 3528–3541, 2019. 3, 6

    2019

  52. [61]

    Zhang, Jingyun Liang, Jiezhang Cao, Bi- han Wen, Radu Timofte, and Luc Van Gool

    Yuanzhi Zhu, K. Zhang, Jingyun Liang, Jiezhang Cao, Bi- han Wen, Radu Timofte, and Luc Van Gool. Denois- ing diffusion models for plug-and-play image restoration. 2023 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition Workshops (CVPRW) , pages 1219–1229,

    2023

  53. [63]

    Use gradient descent to minimize L and update Iˆx

    Image Restoration Algorithm of Fig.2 Given: k = Kernel (size = 9 × 9, σ= 5) Iy = conv (Ix, k) (10) Using existing BIR algorithms to restore Ix: Iˆx0 = BIR (Iy) (11) Design Loss function: L = ∥conv(Iˆx, k) − Iy∥ + λ∥Iˆx − Iˆx0∥ (12) Set Iˆx0 as the initial value of Iˆx. Use gra...

  54. [64]

    Use gradient descent to minimize L and update Iˆx

    Image Restoration Algorithm of Fig.3 Given: k1 = Kernel (size = 13 × 13, σ= 1.0) k2 = Kernel (size = 13 × 13, σ= 1.06, 1.12 or 1.18) Iy = conv (Ix, k1) (13) Design Loss function: L = ∥conv(Iˆx, k2) − Iy∥ + λ∥Iˆx − Ix∥ (14) Set Ix as the initial value of Iˆx. Use gradient desce...

  55. [65]

    Multiple F operations are equivalent to a single operation, F (x, y) = µAx√ 1+σ2 D/σ2 H

    Proof of RGCDI Property Idempotency. Multiple F operations are equivalent to a single operation, F (x, y) = µAx√ 1+σ2 D/σ2 H . Given: µA = COV (y, x) /COV (x, x) nD = y − µAx σ2 D = COV (y, y) − µACOV (y, x) (15) Calculating F (x, F(x, y)): µ′ A = COV (F (x, y), x)/COV (x, x) ...

  56. [66]

    Under this condition, RGCDI P SN Ris bigger than P SN R. Given: x = DW T[Ix] y = DW T[Iy] ˆx = DW T[Iˆx] µA = COV (y, x) /COV (x, x) σ2 D = COV (y, y) − µACOV (y, x) µD = 1/ q 1 + σ2 D/σ2 H µM = arg min µM ∥µM ˆx − µAµDx∥ (21) Calculating ∥IDW T(µM ˆx) − IDW T(µAµDx)∥: (Note: ...

  57. [67]

    Attenuation Net Backbones Selection (De- tailed Experimental Data) Detailed experimental data of attenuation net backbones se- lection experiments are as follows: Table.2,3,4,5,6,7

  58. [68]

    Table.8 shows RACDI test for Classic SR on the BSD100 and Manga109 datasets

    RACDI Testing (Supplementary Experi- mental Results) We apply SCUNet as the Attenuation Net backbone, and test the RACDI on more Datsets. Table.8 shows RACDI test for Classic SR on the BSD100 and Manga109 datasets. Similar to the results on the Urban100 dataset, the max- imum ...

  59. [2023]

    4 CDI: Blind Image Restoration Fidelity Evaluation based on Consistency with Degraded Image Supplementary Material

Pith tools

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