Pith. sign in

REVIEW 6 major objections 7 minor 67 references

Continual Learning-Based Unified Model for Unpaired Image Restoration Tasks

T0 review · 6 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read A single unpaired model can learn to dehaze, desnow, and derain without catastrophic forgetting.

desk verdict A plausible recombination of known components whose central continual-learning claim is not actually demonstrated by the reported experiments; worth referee time only if the authors can supply the missing absolute-performance numbers. read the letter →

arxiv 2507.19184 v1 pith:OCAFFSEB submitted 2025-07-25 cs.CV

classification cs.CV
keywords imagerestorationunpairedlearningcontinualcatastrophicforgettingcycle-consistentGANselectivekernelfusioncycle-contrastivelossadverseweatherremoval
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 argues that a single unpaired image-restoration model can handle three different weather degradations—haze, snow, and rain—learned one after another, without the usual catastrophic forgetting. The authors build on a cycle-consistent GAN for unpaired dehazing and add three mechanisms: selective-kernel fusion layers that adaptively combine multi-scale feature maps, a cycle-contrastive loss that keeps content aligned across domain translation, and elastic weight consolidation (EWC) that protects parameters important to previously learned tasks. On standard benchmarks—RESIDE for dehazing, SRRS for desnowing, and Rain100H for deraining—they report PSNR and SSIM gains over existing unpaired methods, and higher PSNR than several paired methods. The practical payoff the paper seeks is a unified restoration model suitable for applications like autonomous driving, where weather conditions vary and paired training data is scarce.

What carries the argument

The load-bearing machinery is a set of three additions to the AGLC-GAN backbone. Selective Kernel (SK) Fusion layers concatenate a skip-connection feature map and a main-path feature map, compute attention weights via global average pooling followed by a small MLP, and output a weighted sum a1x1 + a2x2, so the network can emphasize the scale most informative for a given degradation. The Cycle-Contrastive Loss samples 64 spatial patches from features extracted at four generator depths, treats the patch at the same location in the reconstructed image as the positive and patches at other locations as negatives, and applies a temperature-scaled cosine-similarity softmax; this pulls corresponding content together while pushing unrelated content apart. Elastic Weight Consolidation adds a Fisher-information-weighted quadratic penalty on parameter deviations from the previously learned solution, which is what lets the model train task 2 and task 3 without erasing task 1. Each component is ablated separately in the paper, and each contributes a PSNR/SSIM improvement on at least one of the three benchmarks.

What would settle it

Train the model with λ chosen on a held-out validation split and with the task order reversed (derain → desnow → dehaze); if the PSNR advantage over AGLC-GAN on RESIDE, SRRS, and Rain100H disappears or forgetting on the first task grows beyond the reported roughly 3–5 dB, the headline results depend on the benchmark-tuned λ=750 schedule rather than on the architecture itself.

Watch

Extended reading notes

Core claim

The paper's central claim is that DA-AGLC-GAN—a cycle-consistent adversarial network formed by adding selective-kernel fusion layers, a cycle-contrastive loss, and elastic weight consolidation to AGLC-GAN—restores hazy, snowy, and rainy images better than the listed unpaired methods, and reports higher PSNR than several paired methods, while the EWC schedule keeps the model from forgetting dehazing when it later learns desnowing and deraining. The single-task experiments report 32.31 dB PSNR / 0.9697 SSIM on RESIDE ITS, 37.13 dB / 0.9793 on RESIDE OTS, 35.53 dB / 0.9432 on SRRS, and 32.23 dB / 0.8434 on Rain100H, each above the AGLC-GAN baseline. The continual-learning experiments sweep the EWC strength λ from 0 to 1000, settle on λ=750, and report forgetting of the earlier task below the no-regularization levels. The claim is therefore twofold: the architectural additions improve per-task restoration quality, and EWC makes the three-task sequence learnable in one model.

Load-bearing premise

The load-bearing premise is that the performance edge is not an artifact of tuning the regularization strength (λ=750) and the task order on the same test benchmarks that produced the headline results.

Editorial extensions

If this is right

  • A single set of weights can serve dehazing, desnowing, and deraining, so a deployed system would not need to detect the weather condition or switch models.
  • Because the method trains on unpaired images, it can in principle be transferred to real-world degraded images where clean/degraded pairs do not exist.
  • The EWC configuration with a stronger penalty on the second consolidation step (λ2=800 vs λ1=700) yields both better final deraining metrics and lower forgetting of the dehazing task, pointing to an asymmetry in how strongly later tasks should be consolidated.
  • On the reported benchmarks, the model exceeds every listed unpaired baseline, showing that unpaired training does not have to concede quality on these restoration tasks.
  • The ablations show each added component—SK fusion and cycle-contrastive loss—improves at least one benchmark over the AGLC-GAN baseline, so the reported gains do not rest on a single ingredient.

Reading between the lines

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

  • A testable extension is to apply the same unpaired continual-learning recipe to other restoration domains, such as low-light enhancement, underwater color correction, or super-resolution, since neither the selective-kernel fusion nor the cycle-contrastive loss is weather-specific.
  • Because the optimal λ was selected on the same benchmarks that produce the headline numbers, the safest reading is that the architecture improves per-task quality while the specific forgetting numbers depend on that tuning; an out-of-sample validation split would settle how much.
  • The paper compares against only a handful of unpaired baselines on the snow and rain benchmarks, so a natural next check is how DA-AGLC-GAN fares against more recent unpaired all-in-one restoration models rather than mainly against the CycleGAN family.
  • If the forgetting asymmetry (consolidate later tasks more strongly) is a general property, task ordering itself becomes a design choice: starting with the hardest or most critical degradation and increasing consolidation strength over time could improve final performance beyond the three-task sequence reported here.
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

6 major / 7 minor

Summary. The manuscript proposes DA-AGLC-GAN, an unpaired CycleGAN-based restoration model intended to handle dehazing, desnowing, and deraining in a single framework. The architecture extends the authors' AGLC-GAN with Selective Kernel Fusion layers, a cycle-contrastive loss, and Elastic Weight Consolidation for sequential task learning. The paper reports single-task results on RESIDE ITS/OTS, SRRS, and Rain100H, compares them with paired and unpaired baselines, and presents an EWC-λ analysis plus component ablations. The central claim is that the proposed model outperforms state-of-the-art unpaired methods and mitigates catastrophic forgetting when tasks are learned sequentially.

Significance. If the claimed results are reproducible, the work would be a useful contribution to unpaired multi-weather image restoration, and the ablation study gives a clear picture of each component's standalone effect. The comparison includes a wide range of paired and unpaired baselines, and the authors are transparent about the EWC parameter analysis. However, the central continual-learning claim is not directly demonstrated: the sequentially trained model's absolute PSNR/SSIM on the three benchmarks is never reported, and the EWC hyperparameters are tuned on the same test sets used for the headline numbers. The absence of error bars, an internal contradiction between Tables IV and VIII, and an ablation outcome in which the full model loses to one of its components on two metrics all make the quantitative conclusions difficult to assess.

major comments (6)
  1. [V-D, Table V, Eq. (18)] The continual-learning experiment reports only forgetting values and λ sweeps; no table or figure gives the final PSNR/SSIM of the sequentially trained DA-AGLC-GAN on the RESIDE ITS/OTS, SRRS, or Rain100H test splits after the deHaze→deSnow→deRain sequence. Since forgetting (Eq. 18) is a relative metric, a model can have low forgetting while remaining poor on all tasks if the EWC penalty prevents adaptation. Please report the absolute final metrics for all three tasks, alongside the single-task results from Tables I–IV, so the unified model's performance can be directly compared.
  2. [V-D, Figs. 7–8, Table V] The EWC regularization strength λ is chosen by sweeping 0–1000 and examining PSNR/SSIM and forgetting on the same benchmark test sets used for the headline results, and the λ1/λ2 comparison in Table V is evaluated on the same test sets. This means the 'optimal' configuration is partly a fitting outcome on the test benchmarks. Please use a held-out validation split or an inner cross-validation, and report variation over multiple seeds or runs.
  3. [V-E, Tables VII–VIII] The ablation tables do not consistently support the claim that DA-AGLC-GAN outperforms its components: on SRRS, AGLC-GAN + SK-Fusion has SSIM 0.9437 versus DA-AGLC-GAN's 0.9432 (Table VII), and on Rain100H, AGLC-GAN + SK-Fusion has PSNR 32.37 versus DA-AGLC-GAN's 32.33 (Table VIII). The statement that DA-AGLC-GAN 'consistently outperforms' the component variants is therefore contradicted; please discuss this outcome and its implications for the fusion design.
  4. [Tables IV and VIII] The same model and dataset, DA-AGLC-GAN on Rain100H, are reported with PSNR 32.23 in Table IV and 32.33 in Table VIII. This internal inconsistency needs to be resolved, and it raises doubt about the reliability of other reported numbers; please verify all benchmark entries against the exact evaluation protocol.
  5. [III-C, Eq. (8)] The cycle-contrastive loss is called 'novel' in the abstract and contributions, but the formulation in Eq. (8), the temperature τ=0.07, and the number of negatives N=64 match CCLGAN [14], which is cited only as motivation. If the loss is identical, please credit CCLGAN and revise the novelty claim; if it differs, specify the differences explicitly.
  6. [V, Tables I–VIII] All reported PSNR/SSIM values appear to be single runs with no error bars or statistical tests. The improvements over the AGLC-GAN baseline are modest in several cases (e.g., 31.69→32.31 on RESIDE ITS, 36.71→37.13 on OTS, 31.79→32.23 on Rain100H), so without variance estimates it is unclear whether these differences are significant. Please report multiple seeds or at least state the run-to-run variability.
minor comments (7)
  1. [Fig. 6] The caption of Figure 6 says 'Results of the proposed model on ITS/OTS dataset' but the figure shows rain removal on Rain100H; please correct the dataset reference.
  2. [IV-B and V-A] Section IV-B reports 60,000 training iterations for all datasets, while Section V-A states 5 epochs with 13k samples for ITS and 1 epoch with 60k samples for OTS; please reconcile these training budgets.
  3. [Abstract] There is a typo, 'dependance', in the abstract; it should be 'dependence'.
  4. [Fig. 8 and V-D] The text says Figure 8 shows final PSNR/SSIM on Task 3 (deRain), but the figure caption says it shows deSnow (Task 2) and deRain (Task 3); please make the description consistent.
  5. [Table V] The header notation 'F-PSNR (ITS T1 →2)' is unclear: forgetting is defined in Eq. (18) as the drop on Task 1 after Tasks 2 and 3, but the table appears to use different pairings. Please define all column abbreviations explicitly.
  6. [Abstract and V] The abstract claims improvements in 'perceptual quality', but the paper reports only PSNR and SSIM; please either include a perceptual metric such as LPIPS or NIQE, or soften the claim.
  7. [References [16] and [27]] Reference [16] is titled 'Cycle-dehaze' but is cited in Section II-B as a CycleGAN baseline, and [27] is the CycleGAN paper; please check that each citation points to the intended work.

Circularity Check

0 steps flagged · score 1.0 of 10

No constructional circularity: the headline results are measured against external benchmarks, and the self-cited AGLC-GAN baseline is an externally published architecture, not an unverified premise.

full rationale

The paper's central quantitative claims are benchmark comparisons on RESIDE ITS/OTS, SRRS, and Rain100H. These tables compare DA-AGLC-GAN against many external methods, so the PSNR/SSIM improvements do not reduce by construction to the method's own definitions. The SK fusion equations (Eqs. 6-7), the cycle-contrastive loss (Eq. 8), and the EWC loss (Eq. 17) are standard formulations imported from cited prior work; none of them defines the reported benchmark numbers. The AGLC-GAN backbone is self-cited (Jaisurya and Mukherjee, 2023) and one coauthor overlaps, but AGLC-GAN is a published, peer-reviewed architecture, and the paper measures its own model against it empirically rather than assuming its superiority. The EWC regularization strength is tuned by sweeping lambda on the same benchmark test sets, which is an experimental-design weakness and could inflate the reported 'optimal' configuration, but this is hyperparameter selection rather than a fitted parameter being renamed as a prediction; no equation in the paper makes the output equal to the tuning criterion by construction. The continual-learning section reports only forgetting metrics and lambda sweeps, not absolute post-sequential PSNR/SSIM on all three test sets, so the unified-model claim is under-supported as evidence, but that is a completeness/correctness issue, not circularity. Finally, the cycle-contrastive loss is explicitly said to be 'motivated by CCLGAN' with the same temperature and negative-sample count, so the abstract's 'novel' label is an attribution concern, not a circular derivation. Overall, the derivation chain is self-contained against external benchmarks and no load-bearing step reduces to its own inputs.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central empirical claims rest on standard deep learning assumptions, on the validity of the chosen benchmarks and metrics, and on several hyperparameters that are either taken from prior work or tuned on the test benchmarks. No new theoretical entities are introduced.

free parameters (5)
  • EWC regularization strength λ = 750
    Selected by sweeping λ from 0 to 1000 and picking the value that minimizes forgetting while retaining final task performance (Section V-D).
  • λ1 and λ2 (EWC penalties after task 1 and task 2) = 700 and 800
    Tuned to balance retention and adaptation; λ2 > λ1 chosen because it gives 'slightly better' deRain performance and less forgetting (Table V).
  • contrastive loss weight λ_contrastive = 0.3
    Set experimentally; no search described (Section III-E).
  • temperature τ in contrastive loss = 0.07
    Taken directly from CCLGAN [14].
  • number of negative samples N = 64
    Set to 64, matching the number of spatial patches.
assumptions (4)
  • domain assumption Unpaired CycleGAN can effectively map degraded to clean domains for haze, snow, and rain
    The backbone AGLC-GAN uses cycle consistency to enable unpaired translation; no proof is given, but it is common practice in the literature.
  • ad hoc to paper Selective Kernel fusion with two specific insertion points improves feature fusion
    The two SK Fusion locations in the generator are chosen without systematic search (Section III-B).
  • domain assumption EWC prevents catastrophic forgetting without needing memory replay
    Relies on prior EWC results [44] and is not formally justified for this setting.
  • domain assumption PSNR and SSIM are sufficient to evaluate restoration quality
    Standard metrics, but they are known to not always align with perceptual quality.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Continual Learning-Based Unified Model for Unpaired Image Restoration Tasks." pith.science (2026). https://pith.science/paper/OCAFFSEB

@misc{pith2026250719184,
  author       = {Pith},
  title        = {Pith review of: Continual Learning-Based Unified Model for Unpaired Image Restoration Tasks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OCAFFSEB}},
  note         = {Machine review of arXiv:2507.19184}
}
read the original abstract

Restoration of images contaminated by different adverse weather conditions such as fog, snow, and rain is a challenging task due to the varying nature of the weather conditions. Most of the existing methods focus on any one particular weather conditions. However, for applications such as autonomous driving, a unified model is necessary to perform restoration of corrupted images due to different weather conditions. We propose a continual learning approach to propose a unified framework for image restoration. The proposed framework integrates three key innovations: (1) Selective Kernel Fusion layers that dynamically combine global and local features for robust adaptive feature selection; (2) Elastic Weight Consolidation (EWC) to enable continual learning and mitigate catastrophic forgetting across multiple restoration tasks; and (3) a novel Cycle-Contrastive Loss that enhances feature discrimination while preserving semantic consistency during domain translation. Further, we propose an unpaired image restoration approach to reduce the dependance of the proposed approach on the training data. Extensive experiments on standard benchmark datasets for dehazing, desnowing and deraining tasks demonstrate significant improvements in PSNR, SSIM, and perceptual quality over the state-of-the-art.

Figures

Figures reproduced from arXiv: 2507.19184 by the authors.

Figure 1
Figure 1. Architecture of the proposed Feature Attention Module. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Architecture of the SK Fusion module. 2) Discriminator: The Discriminator consists of two parts: the Global Discriminator and the Local Discriminator, both having the same architecture. The Global Discriminator pro￾cesses the entire image to capture the global characteristics of the image, while the Local Discriminator processes randomly cut patches from the image to capture local features and variations. The Global… view at source ↗
Figure 3
Figure 3. Overall Architecture of the proposed Generator Model. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Qualitative comparisons of the proposed DA-AGLC-GAN with state-of-the-art (SOTA) models. The first three rows illustrate results on indoor images, [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Results of the proposed model on SRRS dataset. (a) Input Snow [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Results of the proposed model on ITS/OTS dataset. (a) Input rainy [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Variation of forgetting with EWC λ. Lower forgetting values indicate better retention of previously learned tasks. TABLE IV COMPARISONS OF SOTA MODELS OVER RAIN100H DATASET. Model PSNR SSIM Paired Models RESCAN [15] 26.45 0.8458 MCW-Net [61] 30.70 0.922 PReNet [62] 29.…
Figure 8
Figure 8. Figure 8: Final PSNR performance on deSnow (Task 2) and deRain (Task 3) [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Results of the proposed model on all datasets. (a) Input image. (b) Output of AGLC-GAN. (c) Output of AGLC-GAN + SkFusion (d) Output of [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 47 canonical work pages

  1. [14]

    Cycle Contrastive Adversarial Learning for Unsupervised image Deraining

    C. Zhao, W. Cai, C. Hu, and Z. Yuan, “Cycle contrastive adversarial learning for unsupervised image deraining,” 2024. [Online]. Available: https://arxiv.org/abs/2407.11750

  2. [1]

    Single image dehazing using extended local dark channel prior,

    P. Dwivedi and S. Chakraborty, “Single image dehazing using extended local dark channel prior,” Image and vision computing , 2023

  3. [2]

    Ffa-net: Feature fusion attention network for single image dehazing,

    X. Qin, Z. Wang, Y . Bai, X. Xie, and H. Jia, “Ffa-net: Feature fusion attention network for single image dehazing,” in Proceedings of the AAAI conference on artificial intelligence , vol. 34, no. 07, 2020, pp. 11 908–11 915

  4. [3]

    A fast single image haze removal algorithm using color attenuation prior,

    Q. Zhu, J. Mai, and L. Shao, “A fast single image haze removal algorithm using color attenuation prior,” IEEE Transactions on Image Processing , vol. 24, pp. 3522–3533, 2015

  5. [4]

    Non-local image dehazing,

    D. Berman, T. Treibitz, and S. Avidan, “Non-local image dehazing,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2016, pp. 1674–1682

  6. [5]

    A survey on all-in-one image restoration: Taxonomy, evaluation and future trends,

    J. Jiang, Z. Zuo, G. Wu, K. Jiang, and X. Liu, “A survey on all-in-one image restoration: Taxonomy, evaluation and future trends,” 2024. [Online]. Available: https://arxiv.org/abs/2410.15067

  7. [6]

    An introduction to convolutional neural networks,

    K. O’Shea and R. Nash, “An introduction to convolutional neural networks,” arXiv:1511.08458, 2015

  8. [7]

    Review of deep learning: concepts, cnn architectures, challenges, applications, future directions,

    L. Alzubaidi, J. Zhang, A. J. Humaidi, A. Al-Dujaili, Y . Duan, O. Al-Shamma, J. Santamar ´ıa, M. A. Fadhel, M. Al-Amidie, and L. Farhan, “Review of deep learning: concepts, cnn architectures, challenges, applications, future directions,” Journal of big Data , 2021

Show all 67 references
  1. [8]

    Concep- tual understanding of convolutional neural network- a deep learning approach,

    S. Indoliaa, A. K. Goswami, S. P. Mishra, and P. Asopaa, “Concep- tual understanding of convolutional neural network- a deep learning approach,” Procedia computer science , 2018

  2. [9]

    Convolutional neural net- work (cnn) for image detection and recognition,

    R. Chauhan, K. K. Ghanshala, and R. Joshi, “Convolutional neural net- work (cnn) for image detection and recognition,” 2018 first international conference on secure cyber computing and communication (ICSCCC) , 2018

  3. [10]

    Esrgan: Enhanced super-resolution generative adversarial networks,

    X. Wang, K. Yu, S. Wu, J. Gu, Y . Liu, C. Dong, C. C. Loy, Y . Qiao, and X. Tang, “Esrgan: Enhanced super-resolution generative adversarial networks,” European Conference on Computer Vision (ECCV) , 2018

  4. [11]

    Aglc-gan: Attention-based global- local cycle-consistent generative adversarial networks for unpaired single image dehazing,

    R. Jaisurya and S. Mukherjee, “Aglc-gan: Attention-based global- local cycle-consistent generative adversarial networks for unpaired single image dehazing,” Image and Vision Computing , vol. 140, p. 104859, 2023. [Online]. Available: https://www.sciencedirect.com/ science/arti...

  5. [12]

    Attention-based single image dehazing using improved cyclegan,

    R. S. Jaisurya and S. Mukherjee, “Attention-based single image dehazing using improved cyclegan,” in 2022 International Joint Conference on Neural Networks (IJCNN) , 2022, pp. 1–8

  6. [13]

    Single image dehazing using improved cyclegan,

    B. Chaitanya and S. Mukherjee, “Single image dehazing using improved cyclegan,” Journal of Visual Communication and Image Representation, vol. 74, p. 103014, 2021. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S1047320320302248

  7. [15]

    Recurrent squeeze-and- excitation context aggregation net for single image deraining,

    X. Li, J. Wu, Z. Lin, H. Liu, and H. Zha, “Recurrent squeeze-and- excitation context aggregation net for single image deraining,” 2018. [Online]. Available: https://arxiv.org/abs/1807.05698

  8. [16]

    Cycle-dehaze: Enhanced cyclegan for single image dehazing,

    D. Engin, A. Genc, and H. K. Ekenel, “Cycle-dehaze: Enhanced cyclegan for single image dehazing,” in 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , 2018, pp. 938–9388

  9. [17]

    A simple frame- work for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple frame- work for contrastive learning of visual representations,” International conference on machine learning , 2020

  10. [18]

    Unpaired deep image deraining using dual contrastive learning,

    X. Chen, J. Pan, K. Jiang, Y . Li, Y . Huang, C. Kong, L. Dai, and Z. Fan, “Unpaired deep image deraining using dual contrastive learning,” Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022

  11. [19]

    Self-paced weight consolidation for continual learning,

    W. Cong, Y . Cong, G. Sun, Y . Liu, and J. Dong, “Self-paced weight consolidation for continual learning,” IEEE Transactions on Circuits and Systems for Video Technology, 2023

  12. [20]

    A comprehensive survey of continual learning: Theory, method and application,

    L. Wang, X. Zhang, H. Su, and J. Zhu, “A comprehensive survey of continual learning: Theory, method and application,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2024

  13. [21]

    Reinforced continual learning,

    J. Xu and Z. Zhu, “Reinforced continual learning,” Advances in neural information processing systems , 2018

  14. [22]

    Single image haze removal using dark channel prior,

    K. He, J. Sun, and X. Tang, “Single image haze removal using dark channel prior,” IEEE transactions on pattern analysis and machine intelligence, vol. 33, no. 12, pp. 2341–2353, 2010

  15. [23]

    Dehazenet: An end-to-end system for single image haze removal,

    B. Cai, X. Xu, K. Jia, C. Qing, and D. Tao, “Dehazenet: An end-to-end system for single image haze removal,” IEEE Transactions on Image Processing, 2016

  16. [24]

    Draco-dehazenet: An efficient image dehazing network combining de- tail recovery and a novel contrastive learning paradigm,

    G. Y . Leea, T. Damb, M. M. Ferdausc, D. P. Poenara, and V . Duong, “Draco-dehazenet: An efficient image dehazing network combining de- tail recovery and a novel contrastive learning paradigm,” arXiv preprint arXiv:2410.14595, 2024

  17. [25]

    Single image de- hazing algorithm using generative adversarial network based on feature pyramid network,

    X. Cao, J. Dong, S. Zhou, S. Zhao, and D. Zhang, “Single image de- hazing algorithm using generative adversarial network based on feature pyramid network,” International Conference on Computer Vision, Image and Deep Learning (CVIDL) , 2020

  18. [26]

    Wasserstein gan,

    M. Arjovsky, S. Chintala, and L. Bottou, “Wasserstein gan,” 2017. [Online]. Available: https://arxiv.org/abs/1701.07875

  19. [27]

    Unpaired image-to-image translation using cycle-consistent adversarial networks,

    J.-Y . Zhu, T. Park, P. Isola, and A. A. Efros, “Unpaired image-to-image translation using cycle-consistent adversarial networks,” Proceedings of the IEEE international conference on computer vision , 2017

  20. [28]

    Omni-kernel network for image restoration,

    Y . Cui, W. Ren, and A. Knoll, “Omni-kernel network for image restoration,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 2, pp. 1426–1434, Mar. 2024. [Online]. Available: https://ojs.aaai.org/index.php/AAAI/article/view/27907

  21. [29]

    Ucl-dehaze: Toward real-world image dehazing via unsu- pervised contrastive learning,

    Y . Wang, X. Yan, F. L. Wang, H. Xie, W. Yang, X.-P. Zhang, J. Qin, and M. Wei, “Ucl-dehaze: Toward real-world image dehazing via unsu- pervised contrastive learning,” IEEE Transactions on Image Processing, 2024

  22. [30]

    Disentangled bad weather removal gan for pedestrian detection,

    H. Yang, A. Carballo, and K. Takeda, “Disentangled bad weather removal gan for pedestrian detection,” in 2022 IEEE 95th Vehicular Technology Conference: (VTC2022-Spring), 2022, pp. 1–6

  23. [31]

    Addressing domain discrepancy: A dual-branch collaborative model to unsupervised dehazing,

    S. Fan, M. Xue, A. Ning, and S. Zhong, “Addressing domain discrepancy: A dual-branch collaborative model to unsupervised dehazing,” Pattern Recognition Letters , vol. 189, pp. 150–156, 2025. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0167865525000133

  24. [32]

    Pseudo-labeling and confirmation bias in deep semi-supervised learn- ing,

    E. Arazo, D. Ortego, P. Albert, N. E. O’Connor, and K. McGuinness, “Pseudo-labeling and confirmation bias in deep semi-supervised learn- ing,” 2020 International joint conference on neural networks (IJCNN) , 2020. 12

  25. [33]

    Learning from history: Task- agnostic model contrastive learning for image restoration,

    G. Wu, J. Jiang, K. Jiang, and X. Liu, “Learning from history: Task- agnostic model contrastive learning for image restoration,” Proceedings of the AAAI Conference on Artificial Intelligence , 2024

  26. [34]

    Self-supervised models are continual learners,

    E. Fini, V . G. T. da Costa, X. Alameda-Pineda, E. Ricci, K. Alahari, and J. Mairal, “Self-supervised models are continual learners,” 2022. [Online]. Available: https://arxiv.org/abs/2112.04215

  27. [35]

    Continual all-in-one adverse weather removal with knowledge replay on a unified network structure,

    D. Cheng, Y . Ji, D. Gong, Y . Li, N. Wang, J. Han, and D. Zhang, “Continual all-in-one adverse weather removal with knowledge replay on a unified network structure,” 2024. [Online]. Available: https://arxiv.org/abs/2403.07292

  28. [36]

    Few-shot image generation with elastic weight consolidation,

    Y . Li, R. Zhang, J. Lu, and E. Shechtman, “Few-shot image generation with elastic weight consolidation,” arXiv preprint arXiv:2012.02780 , 2020

  29. [37]

    Contrastive learning for compact single image dehazing,

    H. Wu, Y . Qu, S. Lin, J. Zhou, R. Qiao, Z. Zhang, Y . Xie, and L. Ma, “Contrastive learning for compact single image dehazing,” in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, pp. 10 546–10 555

  30. [38]

    Multi-scale context aggregation by dilated convolutions,

    F. Yu, “Multi-scale context aggregation by dilated convolutions,” arXiv preprint arXiv:1511.07122, 2015

  31. [39]

    Deformable convnets v2: More deformable, better results,

    X. Zhu, H. Hu, S. Lin, and J. Dai, “Deformable convnets v2: More deformable, better results,” 2018. [Online]. Available: https://arxiv.org/abs/1811.11168

  32. [40]

    Selective kernel networks,

    X. Li, W. Wang, X. Hu, and J. Yang, “Selective kernel networks,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2019

  33. [41]

    Efficient dehazing with recursive gated convolution in u-net: A novel approach for image dehazing,

    Z. Wang, J. Jia, P. Lyu, and J. Min, “Efficient dehazing with recursive gated convolution in u-net: A novel approach for image dehazing,” Journal of Imaging , vol. 9, p. 183, 09 2023

  34. [42]

    Enhanced pix2pix dehazing network,

    Y . Qu, Y . Chen, J. Huang, and Y . Xie, “Enhanced pix2pix dehazing network,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2019, pp. 8160–8168

  35. [43]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” in International Conference on Learning Representations, 2015

  36. [44]

    Overcoming catastrophic forgetting in neural networks,

    J. Kirkpatrick, R. Pascanu, N. Rabinowitz, J. Veness, G. Desjardins, A. A. Rusu, K. Milan, J. Quan, T. Ramalho, A. Grabska-Barwinska, D. Hassabis, C. Clopath, D. Kumaran, and R. Hadsell, “Overcoming catastrophic forgetting in neural networks,” Proceedings of the national acade...

  37. [45]

    Benchmarking single-image dehazing and beyond,

    B. Li, W. Ren, D. Fu, D. Tao, D. Feng, W. Zeng, and Z. Wang, “Benchmarking single-image dehazing and beyond,” IEEE Transactions on Image Processing , vol. 28, no. 1, pp. 492–505, 2019

  38. [46]

    Jstasr: Joint size and transparency-aware snow removal algorithm based on modified partial convolution and veiling effect removal,

    W.-T. Chen, H.-Y . Fang, J.-J. Ding, C.-C. Tsai, and S.-Y . Kuo, “Jstasr: Joint size and transparency-aware snow removal algorithm based on modified partial convolution and veiling effect removal,” in European Conference on Computer Vision , 2020

  39. [47]

    Joint rain detection and removal from a single image,

    W. Yang, R. T. Tan, J. Feng, J. Liu, Z. Guo, and S. Yan, “Joint rain detection and removal from a single image,” IEEE Conference on Computer Vision and Pattern Recognition , 2017

  40. [48]

    Aod-net: All-in-one dehazing network,

    B. Li, X. Peng, Z. Wang, J. Xu, and D. Feng, “Aod-net: All-in-one dehazing network,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV) , Oct 2017

  41. [49]

    Attention-adaptive multi- scale feature aggregation dehazing network,

    Z. Su, R. Liu, Y . Feng, and F. Zhou, “Attention-adaptive multi- scale feature aggregation dehazing network,” Journal of Visual Communication and Image Representation , vol. 90, p. 103706, 2023. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S1047320322002267

  42. [50]

    Rethinking the elementary function fusion for single-image dehazing,

    Y . Rohn, “Rethinking the elementary function fusion for single-image dehazing,” 2024. [Online]. Available: https://arxiv.org/abs/2405.15817

  43. [51]

    Casdyf-net: Image dehazing via cascaded dynamic filters,

    W. Yinglong and H. Bin, “Casdyf-net: Image dehazing via cascaded dynamic filters,” 2024. [Online]. Available: https://arxiv.org/abs/2409. 08510

  44. [52]

    Self- augmented unpaired image dehazing via density and depth decomposi- tion,

    Y . Yang, C. Wang, R. Liu, L. Zhang, X. Guo, and D. Tao, “Self- augmented unpaired image dehazing via density and depth decomposi- tion,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , June 2022, pp. 2037–2046

  45. [53]

    Unpaired deep image dehazing using contrastive disentanglement learning,

    X. Chen, Z. Fan, P. Li, L. Dai, C. Kong, Z. Zheng, Y . Huang, and Y . Li, “Unpaired deep image dehazing using contrastive disentanglement learning,” in Computer Vision – ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XVII. Berlin,...

  46. [54]

    Unsupervised single image dehazing with generative adversarial network,

    W. Ren, L. Zhou, and J. Chen, “Unsupervised single image dehazing with generative adversarial network,” Multimed. Syst., Feb. 2022

  47. [55]

    Mixdehazenet : Mix structure block for image dehazing network,

    L. Lu, Q. Xiong, D. Chu, and B. Xu, “Mixdehazenet : Mix structure block for image dehazing network,” 2023. [Online]. Available: https://arxiv.org/abs/2305.17654

  48. [56]

    Dehaze-glcgan: Unpaired single image de-hazing via adversarial training,

    Z. Anvari and V . Athitsos, “Dehaze-glcgan: Unpaired single image de-hazing via adversarial training,” 2020. [Online]. Available: https: //arxiv.org/abs/2008.06632

  49. [57]

    Desnownet: Context-aware deep network for snow removal,

    Y .-F. Liu, D.-W. Jaw, S.-C. Huang, and J.-N. Hwang, “Desnownet: Context-aware deep network for snow removal,” IEEE Transactions on Image Processing, vol. 27, no. 6, p. 3064–3073, Jun. 2018. [Online]. Available: http://dx.doi.org/10.1109/TIP.2018.2806202

  50. [58]

    Snow mask guided adaptive residual network for image snow removal,

    B. Cheng, J. Li, Y . Chen, S. Zhang, and T. Zeng, “Snow mask guided adaptive residual network for image snow removal,” 2022. [Online]. Available: https://arxiv.org/abs/2207.04754

  51. [59]

    Lmqformer: A laplace-prior-guided mask query transformer for lightweight snow removal,

    J. Lin, N. Jiang, Z. Zhang, W. Chen, and T. Zhao, “Lmqformer: A laplace-prior-guided mask query transformer for lightweight snow removal,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 33, no. 11, p. 6225–6235, Nov. 2023. [Online]. Available: http://dx.d...

  52. [60]

    Revitalizing convolutional network for image restoration,

    Y . Cui, W. Ren, X. Cao, and A. Knoll, “Revitalizing convolutional network for image restoration,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 46, no. 12, pp. 9423–9438, 2024

  53. [61]

    Mcw-net: Single image deraining with multi-level connections and wide regional non-local blocks,

    Y . Park, M. Jeon, J. Lee, and M. Kang, “Mcw-net: Single image deraining with multi-level connections and wide regional non-local blocks,” Signal Processing: Image Communication, vol. 105, p. 116701, Jul. 2022. [Online]. Available: http://dx.doi.org/10.1016/j.image.2022. 116701

  54. [62]

    Progressive image deraining networks: A better and simpler baseline,

    D. Ren, W. Zuo, Q. Hu, P. Zhu, and D. Meng, “Progressive image deraining networks: A better and simpler baseline,” in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019, pp. 3932–3941

  55. [63]

    Restormer: Efficient transformer for high-resolution image restoration,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, and M.-H. Yang, “Restormer: Efficient transformer for high-resolution image restoration,” 2022. [Online]. Available: https://arxiv.org/abs/2111.09881

  56. [64]

    Image restoration with mean-reverting stochastic differential equations,

    Z. Luo, F. K. Gustafsson, Z. Zhao, J. Sj ¨olund, and T. B. Sch ¨on, “Image restoration with mean-reverting stochastic differential equations,” 2023. [Online]. Available: https://arxiv.org/abs/2301.11699

  57. [65]

    Controlling vision-language models for multi-task image restoration,

    Z. Luo, F. K. Gustafsson, Z. Zhao, J. Sjolund, and T. B. Schon, “Controlling vision-language models for multi-task image restoration,”

  58. [66]

    Semi-supervised transfer learning for image rain removal,

    W. Wei, D. Meng, Q. Zhao, Z. Xu, and Y . Wu, “Semi-supervised transfer learning for image rain removal,” in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , 2019, pp. 3872– 3881. Kotha Kartheek is a senior at the Shiv Nadar Institute of Eminence in ...

  59. [2024]

    Available: https://arxiv.org/abs/2310.01018

    [Online]. Available: https://arxiv.org/abs/2310.01018

Pith tools

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