Pith. sign in

REVIEW 5 major objections 5 minor 73 references

AdaQual-Diff: Diffusion-Based Image Restoration via Adaptive Quality Prompting

T0 review · 5 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read This paper claims that in diffusion-based restoration, prompt length should be set by local perceptual quality: the worse a region scores, the longer and more precise its restoration prompt.

desk verdict Adaptive prompt-length modulation by regional quality scores is a genuinely new and plausible mechanism, but the paper currently can't be verified because its own Table 1 and Section 5.2 quote different numbers for the headline CDD-11 result. read the letter →

arxiv 2504.12605 v1 pith:VKV3LMN7 submitted 2025-04-17 cs.CV

Xin Su , Chen Wu , Yu Zhang , Chen Lyu , Zhuoran Zheng This is my paper · ORCID
classification cs.CV
keywords imagerestorationdiffusionmodelsall-in-onequalityassessmentadaptivepromptingspatiallyvaryingdegradationquality-awareconditioningDeQAScore
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

AdaQual-Diff claims that the right amount of guidance for a diffusion-based restorer is set by local perceptual quality: areas that score badly should receive longer, more detailed restoration prompts, and areas that already look good should get short, preservation-focused prompts. The paper implements this as an explicit inverse mapping from a DeQAScore quality map to prompt complexity, plus a quality-weighted training loss, inside a two-step diffusion model. If the claim holds, it means a single all-in-one model can adapt to mixed, spatially varying degradations without extra parameters or extra sampling iterations. Reported results include 30.11 dB PSNR and 0.9001 SSIM on CDD-11, ahead of the OneRestore baseline, and competitive scores on snow, rain, and raindrop benchmarks.

What carries the argument

The load-bearing mechanism is the Adaptive Quality Prompting algorithm (Algorithm 1) built on the inverse-complexity mapping of Eq. 6. It takes a DeQAScore spatial quality map $Q(y) \in [1,5]^{H\times W}$, partitions the image adaptively, and for each region reduces the map to its mean score $q_r$; that single scalar sets the prompt token count $C_p$, selects a high- or low-quality prompt pool via threshold $\tau$, and picks the top-$C_p$ prompts by feature similarity with the region. This selected prompt set then conditions a two-step diffusion NAFNet, and a quality-weighted loss concentrates training gradients on low-quality regions.

What would settle it

Build a test image whose regions are matched for mean DeQAScore but differ in degradation type and required operation, such as haze versus rain streaks. If an oracle that knows the true degradation per region produces different prompts and better restorations than AdaQual-Diff's quality-only selection, the sufficiency claim fails; a second check is that the reported swing from 28.91 to 30.11 dB between $\tau=1.5$ and $\tau=3.0$ makes the threshold robustness directly measurable on other datasets.

Watch

Extended reading notes

Core claim

The central discovery is a formal relationship between perceptual quality and guidance complexity in diffusion restoration, stated as prompt complexity scaling inversely with local quality: $C_p(q) = C_{\min} + (C_{\max} - C_{\min}) \cdot \left(1 - \frac{q - q_{\min}}{q_{\max} - q_{\min}}\right)$. The authors argue that a quality map from DeQAScore, which scores regions on a 1–5 scale, is not just an evaluation metric but an active conditioning signal. Their Adaptive Quality Prompting partitions the image, reads the mean score per region, computes a prompt length from this formula, and selects that many prompt tokens by feature similarity from a quality-appropriate pool, so severely degraded regions get elaborate, targeted prompts while clean regions get minimal preservation prompts. Combined with a quality-weighted loss that upweights errors in low-quality regions, this yields a spatially varying guidance field inside a diffusion model with a NAFNet backbone, and the paper reports that it outperforms fixed-length prompting and prior all-in-one baselines on composite degradations.

Load-bearing premise

The system assumes that a single mean DeQAScore per region is a sufficient statistic for choosing the right restoration prompt, and that one fitted threshold separates the two prompt pools.

Editorial extensions

If this is right

  • Fixed-length prompts become a strictly weaker design: the ablation table shows adaptive lengths beat C=10, C=20, and C=30 uniformly, so any all-in-one restorer using constant guidance leaves quality on the table.
  • Quality assessment stops being a post-hoc evaluator and becomes part of the generative loop; other restoration frameworks can adopt the same score-to-guidance pattern by swapping in their own quality estimator.
  • Spatial heterogeneity can be handled without extra parameters or extra inference iterations: the same two-step diffusion model allocates effort by prompt selection rather than by running more steps in bad areas.
  • A model trained on composite degradations transfers to single-weather benchmarks, suggesting the quality-driven prompts do not overfit to one degradation type.
  • Downstream tasks benefit: object-detection confidence scores on degraded regions improve after restoration, which matters for autonomous-driving and surveillance use cases.

Reading between the lines

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

  • I infer that the linear inverse mapping is a design choice, not a derived optimum; an experiment comparing Eq. 6 against nonlinear and learned mappings on CDD-11 would show whether the claimed mathematical precision is doing the work or whether any increasing function of $1 - q$ would do.
  • I infer the same complexity-allocation principle could be applied to computation itself—more diffusion steps, larger receptive fields, or more channels in low-quality regions—which would test whether prompt length is the mechanism or just a proxy for attention.
  • I infer that collapsing each region to its mean quality score is the fragile step: a region containing both heavy haze and a sharp object gets an averaged prompt that may neither de-haze fully nor preserve detail, so a distribution-aware or degradation-aware regional descriptor is a natural extension.
  • I infer the caching of quality maps across training iterations assumes the degraded input's quality field is stable; for video or streaming restoration, recomputing spatial quality each frame may be needed, and the cost of DeQAScore would then dominate inference.
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

5 major / 5 minor

Summary. AdaQual-Diff proposes a diffusion-based all-in-one image restoration method in which prompt complexity is modulated by spatially local quality scores. The quality map from DeQAScore is partitioned into regions; for each region, the mean quality is mapped through a linear inverse relation (Eq. 6) to a prompt length, and prompts are selected from high/low pools by feature similarity (Algorithm 1). The training loss adds a quality-weighted noise term and a region-selective perceptual term (Eqs. 7-10). Experiments on CDD-11 and weather benchmarks report 30.11 dB/0.9001 SSIM on CDD-11 (Table 1) and claim state-of-the-art performance on composite degradation and specific weather removal tasks, with 61.12M parameters and two sampling steps. The paper also claims a theoretical connection between perceptual quality and optimal guidance complexity, though the connection is presented through a postulated linear mapping rather than a derived optimality result.

Significance. The central idea of replacing a fixed global prompt with a quality-driven spatially varying prompt, and weighting the loss by local quality, is timely and, if implemented as described, has interesting practical value. The ablation tables (Tables 3-5, supplementary Table 8) are internally consistent and show a clear advantage of the adaptive scheme over fixed-length prompts, and the reported efficiency (two sampling steps, 17 ms per image) is attractive. The paper's contributions are weakened, however, by an unreconciled numerical discrepancy in the headline CDD-11 result, an overbroad state-of-the-art claim on weather removal, and the absence of a derivation that would justify the 'theoretical framework' language. These issues must be resolved before the experimental claims can be fully accepted.

major comments (5)
  1. [§5.2 and Table 1] The text states that AdaQual-Diff 'outperforms the previous best method by a large margin with 31.02 dB PSNR and 0.9091 SSIM,' while Table 1 reports 30.11 dB/0.9001 SSIM for AdaQual-Diff and 28.72/0.8821 for OneRestore†. The two versions of the headline result differ by 0.91 dB and 0.009 SSIM, and the claimed margin over OneRestore† changes from roughly 2.3 dB to 1.4 dB depending on which number is correct. Please identify the correct set of numbers and reconcile all occurrences; as written, the central quantitative claim is ambiguous.
  2. [§5.2 and Table 2] The claim that AdaQual-Diff 'achieves state-of-the-art performance on specific weather removal tasks' is not supported by Table 2: on Outdoor-Rain the PSNR is 31.81, below T3-DiffWeather's 31.99, and on RainDrop the SSIM is 0.9330, below T3-DiffWeather's 0.9411. Please qualify the claim to name the datasets and metrics on which the method is actually best, or present the comparison with a more precise scope.
  3. [§1, Eq. (6), Supplementary Table 8] The contribution statement says the paper 'establishes a theoretical framework that formally connects perceptual quality assessment with optimal guidance complexity,' but the paper does not derive Eq. (6) from an optimality criterion; it postulates a linear inverse mapping between quality and prompt complexity. The free parameters C_min, C_max, tau, lambda1, and lambda2 are selected by ablations on CDD-11 validation, and supplementary Table 8 shows a 1.2 dB swing between tau=1.5 and tau=4.5. Please either provide a formal derivation or justification for the inverse linear relation, or rephrase the contribution as a design heuristic, and report how the selected tau transfers to the other benchmarks rather than only to CDD-11.
  4. [Algorithm 1] Algorithm 1 reduces each partition region to its mean quality q_r (step 7) and selects prompt content only through the magnitude C_p and a binary high/low pool decision (step 9), discarding all spatial structure and degradation identity within a region. If two regions with the same mean score but different degradation types require different prompt content, this mechanism cannot deliver the 'precise restoration directives' claimed. Please add an experiment or analysis that tests whether the mean statistic suffices, for example by comparing against per-pixel quality-conditioned prompting or by reporting the frequency with which regions mix different degradation types.
  5. [Availability] No code, model weights, or evaluation scripts are provided, so the conflicting CDD-11 numbers and the weather benchmark results cannot be independently checked. I recommend releasing these artifacts with the revision, at minimum for the CDD-11 comparison, so that the reported PSNR/SSIM values can be reproduced by readers.
minor comments (5)
  1. [§5.3] The text reads 'across different different adverse weather conditions'; the duplicated word should be removed.
  2. [Fig. 1 caption and §4] The wording 'Our analysis reveals a mathematical relationship' overstates what is in fact a postulated linear mapping; consider softening this to 'we hypothesize and implement a relationship' until a derivation is supplied.
  3. [Table 5] The row labels are ambiguous: the column headers say L1, Prompt, and Quality-weighted, but the checkmarks do not make clear which row corresponds to 'Full Loss (Ours)' other than all boxes being checked; please make the rows self-explanatory.
  4. [Eq. (9) and §3] Eq. (9) uses w(q)=(q_max-q)/(q_max-q_min), which is zero at q_max, and DeQAScore is described as producing values in [1,5]; please clarify whether q_min and q_max in Eq. (6) and Eq. (9) are fixed dataset bounds (1 and 5) or are recomputed per image.
  5. [Supplementary Section B] The supplementary text states that the RainDrop-A test subset contains 58 images, while the main paper reports results on 'RainDrop' without specifying the subset; please align the terminology and specify the exact test split used.

Circularity Check

1 steps flagged · score 5.0 of 10

Inverse-scaling 'theory' is definitional (Eq. 6), but benchmarks are external: partial circularity.

  1. self definitional [Section 1 (contribution bullet); Section 4, Eq. (6); Figure 1 caption; Algorithm 1 step 8]
    "We establish a theoretical framework that formally connects perceptual quality assessment with optimal guidance complexity in diffusion models, demonstrating that prompt complexity should scale inversely with local image quality to maximize restoration efficacy. [Also:] The core technical contribution is our Adaptive Quality Prompting mechanism, which modulates prompt complexity C_p according to local quality scores: C_p(q)=C_min+(C_max-C_min)·(1 - (q-q_min)/(q_max-q_min)), (6)."

    The claimed 'demonstration' is not derived from diffusion theory or DeQAScore; Eq. (6) defines prompt complexity as a decreasing affine function of regional quality, and Algorithm 1 applies that definition directly. Thus the inverse relationship is true by construction inside the model, not an independent result. The ablations in Table 3 compare adaptive versus fixed prompt lengths and can support adaptivity, but they never vary the direction of the quality-to-length mapping, so they cannot independently confirm the asserted directional law. The paper's central conceptual 'prediction' therefore reduces to its own input definition.

full rationale

The paper's strongest conceptual claim is partially circular: Section 1 presents as a theoretical framework a relationship that is actually installed by definition in Eq. (6), so the inverse-complexity principle is a design choice rather than an independently derived result. However, the quantitative contribution is not forced by that definition. Table 1 and Table 2 compare against published methods on public benchmarks, and Tables 3 and 4 confront the adaptive design with fixed-prompt and alternative-quality-estimator baselines, giving the restoration results independent empirical content. DeQAScore [56] is from non-overlapping authors, so no load-bearing self-citation is involved. Separately, the text's internal inconsistency (Section 5.2 reports 31.02 dB / 0.9091 SSIM while Table 1 reports 30.11 dB / 0.9001 SSIM) and the absence of released code are correctness and reproducibility concerns, not circularity. The supplementary limitation section (F) acknowledges computational cost and parameter count but does not allege any circular step.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The mechanism rests on four premises: DeQAScore maps are trustworthy spatial quality estimates (cited, not re-validated); the inverse-linear quality-to-complexity mapping is near-optimal (asserted, not derived); token count controls restoration strength (partially ablated); and prompt retrieval via F_e similarity delivers the right content (unspecified). Several free design choices (C_min, C_max, τ, λ1, λ2, w(q)) are fitted on CDD-11 validation, the same benchmark family used for the headline result. There are no invented entities; the contribution is a new control policy over a known prompt-pool diffusion architecture.

free parameters (4)
  • C_min, C_max (prompt complexity bounds) = not reported
    Eq. 6's inverse linear mapping; the paper never gives values or a justification beyond 'bounds.'
  • τ (quality threshold) = 3.0
    Selected by ablation on CDD-11 (supplementary Table 8); performance drops from 30.11 to 28.91 at τ=1.5 and 29.75 at τ=4.5, so the headline result depends on this fitted value.
  • λ1, λ2 (loss weights) = λ1=0.5, λ2=0.1
    Chosen by 'ablation studies on validation data' (Sec 4.2), i.e., fitted to the benchmark used for the headline numbers.
  • w(q) quality weighting = (q_max-q)/(q_max-q_min)
    The loss weight in Eq. 9 has the same functional form as the prompt complexity in Eq. 6, so the two components are not independent tests of the inverse-quality hypothesis.
assumptions (4)
  • domain assumption DeQAScore produces spatial quality maps in [1,5] that reliably localize degradation severity and are available for new images at inference.
    Algorithm 1 and Eq. 5 depend entirely on this map; the paper cites DeQAScore [56] but does not validate its spatial accuracy on CDD-11, Snow100K, Outdoor-Rain, or RainDrop.
  • ad hoc to paper Optimal guidance complexity is monotonically decreasing in regional quality, and the linear form of Eq. 6 suffices.
    Figure 1 and the Introduction assert C_p ∝ f(1-Q) and 'optimal guidance complexity' without specifying f or proving optimality; Eq. 6 is the definition, not a derived result.
  • domain assumption Prompt token count is a valid, controllable axis of restoration guidance strength in the prompt-pool diffusion framework.
    Algorithm 1 step 10 assumes that selecting more prompts (larger C_p) monotonically strengthens restoration; Table 3 shows C=30 > C=10 globally, but does not establish local monotonic control over generation quality.
  • domain assumption The region embedding F_e(r) and prompt pool similarity metric select prompts containing the right restoration content.
    Algorithm 1 step 10 uses TopK(Similarity(P_pool, F_e(r))) but neither the embedding F_e nor the similarity function is described, so the content-selection mechanism is unspecified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AdaQual-Diff: Diffusion-Based Image Restoration via Adaptive Quality Prompting." pith.science (2026). https://pith.science/paper/VKV3LMN7

@misc{pith2026250412605,
  author       = {Pith},
  title        = {Pith review of: AdaQual-Diff: Diffusion-Based Image Restoration via Adaptive Quality Prompting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VKV3LMN7}},
  note         = {Machine review of arXiv:2504.12605}
}
read the original abstract

Restoring images afflicted by complex real-world degradations remains challenging, as conventional methods often fail to adapt to the unique mixture and severity of artifacts present. This stems from a reliance on indirect cues which poorly capture the true perceptual quality deficit. To address this fundamental limitation, we introduce AdaQual-Diff, a diffusion-based framework that integrates perceptual quality assessment directly into the generative restoration process. Our approach establishes a mathematical relationship between regional quality scores from DeQAScore and optimal guidance complexity, implemented through an Adaptive Quality Prompting mechanism. This mechanism systematically modulates prompt structure according to measured degradation severity: regions with lower perceptual quality receive computationally intensive, structurally complex prompts with precise restoration directives, while higher quality regions receive minimal prompts focused on preservation rather than intervention. The technical core of our method lies in the dynamic allocation of computational resources proportional to degradation severity, creating a spatially-varying guidance field that directs the diffusion process with mathematical precision. By combining this quality-guided approach with content-specific conditioning, our framework achieves fine-grained control over regional restoration intensity without requiring additional parameters or inference iterations. Experimental results demonstrate that AdaQual-Diff achieves visually superior restorations across diverse synthetic and real-world datasets.

Figures

Figures reproduced from arXiv: 2504.12605 by the authors.

Figure 1
Figure 1. Our analysis reveals a mathematical relationship [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our AdaQual-Diff framework. Left: The complete pipeline incorporates region-wise quality assessment [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Quality assessment visualization. (a) An example [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Comparison of image restoration on low+haze+rain (the upper two examples) and low+haze+snow (the lower two [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Comparison of image restoration under real-world adverse weather conditions. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Adaptive Quality Prompting analysis. (a) Input. (b [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Object detection performance in adverse weather [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Structure of the NAFNet Block. D ADDITIONAL QUALITATIVE RESULTS D.1 More Results on CDD-11 Dataset We provide additional visual comparisons on the CDD-11 dataset to demonstrate the effectiveness of AdaQual-Diff in handling complex composite degradations. Figures 9 and …
Figure 9
Figure 9. Figure 9: Qualitative comparison on the CDD-11 dataset. [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Qualitative comparison on the CDD-11 dataset. [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: Qualitative comparisons on benchmark datasets Snow100K, Outdoor-Rain, and RainDrop. [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: Abl. of the quality threshold 𝜏 [PITH_FULL_IMAGE:figures/full_fig_p013_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

73 extracted references · 58 canonical work pages

  1. [1]

    Amir Bar, Yossi Gandelsman, Trevor Darrell, Amir Globerson, and Alexei Efros

  2. [2]

    Sebastian Bosse, Dominique Maniry, Klaus-Robert Muller, Thomas Wiegand, and Wojciech Samek. 2018. Deep neural networks for no-reference and full-reference image quality assessment. IEEE TIP 27, 1 (2018), 206–219

    work page 2018

  3. [3]

    Liangyu Chen, Xiaojie Chu, Xiangyu Zhang, and Jian Sun. 2022. Simple baselines for image restoration. In European Conference on Computer Vision . Springer, 17–33

    work page 2022

  4. [4]

    Sixiang Chen, Tian Wu, Junshi Li, and Rui Li. 2023. UDR-S2Former: Uncertainty- Driven Raindrop Removal with Vision Transformers. In Proceedings of the IEEE International Conference on Computer Vision (ICCV) . 5678–5687

    work page 2023

  5. [5]

    Sixiang Chen, Tian Ye, Jinbin Bai, Erkang Chen, Jun Shi, and Lei Zhu. 2023. Sparse sampling transformer with uncertainty-driven ranking for unified removal of raindrops and rain streaks. InProceedings of the IEEE/CVF International Conference on Computer Vision. 13106–13117

    work page 2023

  6. [6]

    Sixiang Chen, Tian Ye, Zhaohu Zhang, Kaiand Xing, Yunlong Lin, and Lei Zhu

    work page

  7. [7]

    Wei-Ting Chen, Hao-Yu Fang, Jian-Jiun Ding, Cheng-Che Tsai, and Sy-Yen Kuo

    work page

  8. [8]

    Wei-Ting Chen, Zhi-Kai Huang, Cheng-Che Tsai, Hao-Hsiang Yang, Jian-Jiun Ding, and Sy-Yen Kuo. 2022. Learning multiple adverse weather removal via two-stage knowledge learning and multi-contrastive regularization: Toward a unified model. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 17653–17662

    work page 2022

Show all 73 references
  1. [9]

    Yuning Cui, Wenqi Ren, and Alois Knoll. 2024. Omni-Kernel Network for Image Restoration. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38. 1426–1434

    2024

  2. [10]

    Daocheng Fu, Xin Li, Licheng Wen, Min Dou, Pinlong Cai, Botian Shi, and Yu Qiao

  3. [11]

    Arindam Ghosh, Samuel Daulton, and Mihaela van der Schaar. 2024. On the Expressive Power of Normalizing Flows and Variational Autoencoders. In The Twelfth International Conference on Learning Representations (ICLR)

    2024

  4. [12]

    Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde- Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio

    Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde- Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2014. Generative Adversarial Networks. In Advances in Neural Information Processing Systems (NeurIPS), Vol. 27. 2672–2680

    2014

  5. [13]

    Jinjin Gu, Haoming Cai, Haoyu Chen, Xiaoxing Ye, Jimmy Ren, and Chao Dong

  6. [14]

    Lanqing Guo, Chong Wang, Wenhan Yang, Siyu Huang, Yufei Wang, Hanspeter Pfister, and Bihan Wen. 2023. Shadowdiffusion: When degradation prior meets diffusion model for shadow removal. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 14049–14058

    2023

  7. [15]

    Yu Guo, Yuan Gao, Yuxu Lu, Ryan Wen Liu, and Shengfeng He. 2024. OneRestore: A Universal Restoration Framework for Composite Degradation. In European Conference on Computer Vision

    2024

  8. [16]

    Kaiming He, Jian Sun, and Xiaoou Tang. 2010. Single Image Haze Removal Using Dark Channel Prior. IEEE Transactions on Pattern Analysis and Machine Intelligence 33, 12 (2010), 2341–2353

    2010

  9. [17]

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. 2020. Denoising Diffusion Probabilistic Models. Advances in Neural Information Processing Systems (NeurIPS) 33 (2020), 6840–6851. 2020

    2020

  10. [18]

    In European Conference on Computer Vision

    Image quality assessment for perceptual image restoration: A new dataset, benchmark and metric. In European Conference on Computer Vision . Springer, 137–153

  11. [19]

    Priyank Jaini, Andre Fu, Max Z Chen, and Didrik Nielsen. 2024. Multi-Flow VAE: Continuous Normalizing Flows for VAEs Revisited. In The Twelfth International Conference on Learning Representations (ICLR)

    2024

  12. [20]

    Kui Jiang, Zhongyuan Wang, Peng Yi, Chen Chen, Zheng Wang, Xiao Wang, Junjun Jiang, and Chia-Wen Lin. 2021. Rain-free and residue hand-in-hand: A progressive coupled network for real-time image deraining. IEEE Transactions on Image Processing 30 (2021), 7404–7418

    2021

  13. [21]

    Yeying Jin, Wei Ye, Wenhan Yang, Yuan Yuan, and Robby T Tan. 2024. Des3: Adaptive attention-driven self and soft shadow removal using vit similarity. In Proceedings of the AAAI Conference on Artificial Intelligence . 2634–2642

    2024

  14. [22]

    Junjie Ke, Qifei Wang, Yilin Wang, Peyman Milanfar, and Feng Yang. 2021. MUSIQ: Multi-scale image quality transformer. InProceedings of the IEEE/CVF International Conference on Computer Vision . 5148–5157

    2021

  15. [23]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. 2017. Image-to- image translation with conditional adversarial networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . 1125–1134

    2017

  16. [24]

    Haoying Li, Yifan Yang, Meng Chang, Shiqi Chen, Huajun Feng, Zhihai Xu, Qi Li, and Yueting Chen. 2022. Srdiff: Single image super-resolution with diffusion probabilistic models. Neurocomputing 479 (2022), 47–59

    2022

  17. [25]

    Ruoteng Li, Loong-Fah Cheong, and Robby T. Tan. 2019. Heavy rain image restoration: Integrating physics model and conditional adversarial learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    2019

  18. [26]

    Tan, and Loong-Fah Cheong

    Ruoteng Li, Robby T. Tan, and Loong-Fah Cheong. 2020. All in One Bad Weather Removal Using Architectural Search. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . 3175–3185

    2020

  19. [27]

    Xia Li, Jianlong Wu, Zhouchen Lin, Hong Liu, and Hongbin Zha. 2018. Recurrent squeeze-and-excitation context aggregation net for single image deraining. In Proceedings of the European Conference on Computer Vision (ECCV) . 254–269

    2018

  20. [28]

    Boyun Li, Xiao Liu, Peng Hu, Zhongqin Wu, Jiancheng Lv, and Xi Peng. 2022. All-In-One Image Restoration for Unknown Corruption. In IEEE Conference on Computer Vision and Pattern Recognition . New Orleans, LA

    2022

  21. [29]

    Yun-Fu Liu, Da-Wei Jaw, Shih-Chia Huang, and Jenq-Neng Hwang. 2018. Desnownet: Context-aware deep network for snow removal. IEEE Transactions on Image Processing 27, 6 (2018), 3064–3073

    2018

  22. [30]

    Anish Mittal, Anush Krishna Moorthy, and Alan Conrad Bovik. 2012. No- reference image quality assessment in the spatial domain. IEEE TIP 21, 12 (2012), 4695–4708

    2012

  23. [31]

    Chong Mou, Qian Wang, and Jian Zhang. 2022. Deep Generalized Unfolding Networks for Image Restoration. In CVPR

    2022

  24. [32]

    Ozan Özdenizci and Robert Legenstein. 2023. Restoring Vision in Adverse Weather Conditions with Patch-Based Denoising Diffusion Models. IEEE Trans- actions on Pattern Analysis and Machine Intelligence (TPAMI) (2023). MM’25, October 27–31, 2025, Dublin, Ireland Xin Su, Chen Wu,...

    2023

  25. [33]

    Xing Liu, Masanori Suganuma, Zhun Sun, and Takayuki Okatani. 2019. Dual residual networks leveraging the potential of paired operations for image restora- tion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 7007–7016

    2019

  26. [34]

    Vaishnav Potlapalli, Syed Waqas Zamir, Salman Khan, and Fahad Khan. 2023. PromptIR: Prompting for All-in-One Image Restoration. In Thirty-seventh Con- ference on Neural Information Processing Systems

    2023

  27. [35]

    Rui Qian, Robby T Tan, Wenhan Yang, Jiajun Su, and Jiaying Liu. 2018. Attentive generative adversarial network for raindrop removal from a single image. In Proceedings of the IEEE conference on computer vision and pattern recognition . 2482–2491

    2018

  28. [36]

    Ruijie Quan, Xin Yu, Yuanzhi Liang, and Yi Yang. 2021. Removing raindrops and rain streaks in one go. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 9147–9156

    2021

  29. [37]

    Yuhui Quan, Shijie Deng, Yixin Chen, and Hui Ji. 2019. Deep learning for seeing through window with raindrops. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 2463–2471

    2019

  30. [38]

    Ozan Özdenizci and Robert Legenstein. 2023. Restoring vision in adverse weather conditions with patch-based denoising diffusion models. IEEE Transactions on Pattern Analysis and Machine Intelligence (2023), 1–12. https://doi.org/10.1109/ TPAMI.2023.3238179

    2023

  31. [39]

    Binbin Song, Xiangyu Chen, Shuning Xu, and Jiantao Zhou. 2023. Under-display camera image restoration with scattering effect. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 12580–12589

    2023

  32. [40]

    Yuda Song, Zhuqing He, Hui Qian, and Xin Yu. 2023. DehazeFormer: Vision Transformer-Based Architecture for Single Image Dehazing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . 12345–12354

    2023

  33. [41]

    Shaolin Su, Qingsen Yan, Yu Zhu, Cheng Zhang, Xin Ge, Jinqiu Sun, and Yanning Zhang. 2020. Blindly assess image quality in the wild guided by a self-adaptive hyper network. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 3667–3676

    2020

  34. [42]

    Jeya Maria Jose Valanarasu, Rajeev Yasarla, and Vishal M Patel. 2022. Tran- sweather: Transformer-based restoration of images degraded by adverse weather conditions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2353–2363

    2022

  35. [43]

    Chitwan Saharia, Jonathan Ho, William Chan, Tim Salimans, David J Fleet, and Mohammad Norouzi. 2022. Image super-resolution via iterative refinement. IEEE Transactions on Pattern Analysis and Machine Intelligence 45, 4 (2022), 4713–4726

    2022

  36. [44]

    Tianyu Wang, Xin Yang, Ke Xu, Shaozhe Chen, Qiang Zhang, and Rynson WH Lau. 2019. Spatial attentive single-image deraining with a high quality real rain dataset. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 12270–12279

    2019

  37. [45]

    Xinlong Wang, Wen Wang, Yue Cao, Chunhua Shen, and Tiejun Huang. 2023. 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. 6830–6839

    2023

  38. [46]

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. 2004. Image quality assessment: from error visibility to structural similarity. IEEE TIP 13, 4 (2004), 600–612

    2004

  39. [47]

    Haoning Wu, Zicheng Zhang, Erli Zhang, Chaofeng Chen, Liang Liao, Annan Wang, Li Chunyi, Wenxiu Sun, Qiong Yan, Guangtao Zhai, et al. 2024. Q-Bench: A benchmark for general-purpose foundation models on low-level vision. In ICLR (2024)

    2024

  40. [48]

    Jianyi Wang, Zongsheng Yue, Shangchen Zhou, Kelvin C. K. Chan, and Chen Change Loy. 2023. Exploiting Diffusion Prior for Real-World Image Super- Resolution. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). 12345–12354

    2023

  41. [49]

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Chunyi Li, Liang Liao, Annan Wang, Erli Zhang, Wenxiu Sun, Qiong Yan, Xiongkuo Min, Guangtai Zhai, and Weisi Lin. 2023. Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels. arXiv preprint arXiv:2312...

    2023 arXiv

  42. [50]

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Liang Liao, Chunyi Li, Yixuan Gao, Annan Wang, Erli Zhang, Wenxiu Sun, et al . 2024. Q-Align: Teaching LMMs for visual scoring via discrete text-defined levels. In ICML

    2024

  43. [51]

    Jie Xiao, Xueyang Fu, Aiping Liu, Feng Wu, and Zheng-Jun Zha. 2022. Image de-raining transformer. IEEE Transactions on Pattern Analysis and Machine Intelligence (2022)

    2022

  44. [52]

    Zhenhua Xu, Yujia Zhang, Enze Xie, Zhen Zhao, Yong Guo, Kwan-Yee K Wong, Zhenguo Li, and Hengshuang Zhao. 2024. Drivegpt4: Interpretable end-to-end autonomous driving via large language model. IEEE Robotics and Automation Letters (2024)

    2024

  45. [53]

    Haoning Wu, Zicheng Zhang, Erli Zhang, Chaofeng Chen, Liang Liao, Annan Wang, Kaixin Xu, Chunyi Li, Jingwen Hou, Guangtao Zhai, et al. 2024. Q-Instruct: Improving low-level visual abilities for multi-modality foundation models. In Proceedings of the IEEE/CVF Conference on Comp...

    2024

  46. [54]

    Tan, Jiashi Feng, Jiaying Liu, Zongming Guo, and Shuicheng Yan

    Wenhan Yang, Robby T. Tan, Jiashi Feng, Jiaying Liu, Zongming Guo, and Shuicheng Yan. 2017. Deep Joint Rain Detection and Removal from a Single Image. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 1357–1366

    2017

  47. [55]

    Tian Ye, Sixiang Chen, Jinbin Bai, Jun Shi, Chenghao Xue, Jingxia Jiang, Junjie Yin, Erkang Chen, and Yun Liu. 2023. Adverse weather removal with codebook priors. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 12653–12664

    2023

  48. [56]

    Zhiyuan You, Xin Cai, Jinjin Gu, Tianfan Xue, and Chao Dong. 2025. Teaching Large Language Models to Regress Accurate Image Quality Scores using Score Distribution. In IEEE Conference on Computer Vision and Pattern Recognition

    2025

  49. [57]

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, and Ming-Hsuan Yang. 2022. Restormer: Efficient Transformer for High- Resolution Image Restoration. In CVPR

    2022

  50. [58]

    Tao Yang, Peiran Ren, Xuansong Xie, and Lei Zhang. 2023. Pixel-aware stable diffusion for realistic image super-resolution and personalized stylization. arXiv preprint arXiv:2308.14469 (2023)

    2023 arXiv

  51. [59]

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, and Ling Shao. 2021. Multi-stage progressive image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 14821–14831

    2021

  52. [60]

    Kaihao Zhang, Rongqing Li, Yanjiang Yu, Wenhan Luo, and Changsheng Li

  53. [61]

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang

  54. [62]

    Man Zhou, Jie Huang, Chun-Le Guo, and Chongyi Li. 2023. Fourmer: An efficient global modeling paradigm for image restoration. In International conference on machine learning. PMLR, 42589–42601

    2023

  55. [63]

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, and Ling Shao. 2021. Multi-Stage Progressive Image Restoration. In CVPR

    2021

  56. [64]

    Hanwei Zhu, Haoning Wu, Yixuan Li, Zicheng Zhang, Baoliang Chen, Lingyu Zhu, Yuming Fang, Guangtao Zhai, Weisi Lin, and Shiqi Wang. 2024. Adaptive Image Quality Assessment via Teaching Large Multimodal Model to Compare. arXiv preprint arXiv:2405.19298 (2024)

    2024 arXiv

  57. [65]

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. 2017. Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks. In Proceedings of the IEEE International Conference on Computer Vision (ICCV) . 2223– 2232

    2017

  58. [66]

    all-in-one

    Yurui Zhu, Tianyu Wang, Xueyang Fu, Xuanyu Yang, Xin Guo, Jifeng Dai, Yu Qiao, and Xiaowei Hu. 2023. Learning Weather-General and Weather-Specific Features for Image Restoration Under Multiple Adverse Weather Conditions. In Proc. IEEE/CVF Conference on Computer Vision and Patt...

    2023

  59. [70]

    Xiaoyu Zhou, Zhiwei Lin, Xiaojun Shan, Yongtao Wang, Deqing Sun, and Ming- Hsuan Yang. 2024. Drivinggaussian: Composite gaussian splatting for surround- ing dynamic autonomous driving scenes. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognitio...

    2024

  60. [2018]

    Proceedings of the IEEE conference on computer vision and pattern recognition (2018), 586–595

    The unreasonable effectiveness of deep features as a perceptual metric. Proceedings of the IEEE conference on computer vision and pattern recognition (2018), 586–595

    2018

  61. [2020]

    InComputer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXI 16

    Jstasr: Joint size and transparency-aware snow removal algorithm based on modified partial convolution and veiling effect removal. InComputer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXI 16. Springer, 754–770

    2020

  62. [2021]

    IEEE Transactions on Image Processing (2021)

    Deep Dense Multi-scale Network for Snow Removal Using Semantic and Geometric Priors. IEEE Transactions on Image Processing (2021)

    2021

  63. [2022]

    Advances in Neural Information Processing Systems 35 (2022), 25005–25017

    Visual prompting via image inpainting. Advances in Neural Information Processing Systems 35 (2022), 25005–25017

    2022

  64. [2024]

    In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision

    Drive like a human: Rethinking autonomous driving with large language models. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 910–919

  65. [2025]

    Teaching Tailored to Talent: Adverse Weather Restoration via Prompt Pool and Depth-Anything Constraint. In ECCV. 95–115

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.