pith. sign in

arxiv: 2606.18644 · v1 · pith:TWF5U7FQnew · submitted 2026-06-17 · 💻 cs.CV

Spiking Pyramid Wavelet Transformation for High-efficient and Low-energy Image Restoration

Chen Zhao , Xiantao Hu , Song Wu , Qian Wang , Chen Wu , Rui Xie , Jian Yang , Ying Tai This is my paper

Pith reviewed 2026-06-26 21:51 UTC · model grok-4.3

classification 💻 cs.CV
keywords spiking neural networksimage restorationwavelet transformationpyramid modelenergy efficiencycomputational efficiencycomputer vision
0
0 comments X

The pith

A spiking neural network with pyramid wavelet blocks performs image restoration at lower computational cost and energy use while keeping output quality.

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

The paper introduces the spiking pyramid wavelet model (SPWM) to handle image restoration with spiking neural networks. It develops the spiking dual pyramid wavelet block to overcome the limited receptive fields of standard spiking CNN operations by shifting dependency modeling into the wavelet domain. The approach targets efficiency gains for resource-constrained settings. A reader would care if the method delivers comparable restoration results with measurably lower energy draw on standard benchmarks. The reported experiments indicate that quality holds while costs and consumption drop.

Core claim

The central claim is that the SPWM model, centered on the spiking dual pyramid wavelet block, models long-range dependencies and exploits image degradation properties directly in the wavelet domain, enabling spiking neural networks to achieve image restoration with substantially reduced computational costs and energy consumption while preserving quality on multiple benchmarks.

What carries the argument

The spiking dual pyramid wavelet (SDPW) block, which performs dependency modeling and degradation analysis inside the wavelet domain for spiking networks.

If this is right

  • SPWM produces image restoration results on benchmarks at lower computational cost than prior spiking CNN approaches.
  • Energy consumption drops significantly while output quality stays comparable.
  • Spiking networks become more practical for image restoration on devices with tight power budgets.
  • Wavelet-domain processing offers a route to address receptive-field limits in other spiking vision models.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same block structure could be tested on related tasks such as denoising or super-resolution where degradation modeling matters.
  • Hardware measurements on neuromorphic chips would give a clearer picture of real energy savings beyond simulation counts.
  • Combining the pyramid levels with additional spiking mechanisms might further reduce operations without new architectural changes.

Load-bearing premise

The spiking dual pyramid wavelet block can capture long-range dependencies and degradation properties in the wavelet domain without incurring major quality losses.

What would settle it

Running the SPWM model on the same benchmarks and observing either higher restoration error or no reduction in energy or operations compared with baseline spiking CNNs would falsify the claim.

Figures

Figures reproduced from arXiv: 2606.18644 by Chen Wu, Chen Zhao, Jian Yang, Qian Wang, Rui Xie, Song Wu, Xiantao Hu, Ying Tai.

Figure 1
Figure 1. Figure 1: Our motivations. The degraded images can be decomposed into low-frequency sub-image ( [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The effective receptive field visualization. lowest in the LL frequency domain component, and the SSIM is lowest in the HL fre￾quency domain component, illustrating that the degradation is primarily concentrated in specific frequency components. To further achieve long-range dependency model￾ing, we employ a random shuffling (RS) operation in the wavelet domain [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overall architecture of our proposed spiking pyramid wavelet-based model (SPWM). The core [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The structure of the proposed SDPW block. The figure displays the network details in the 2-level [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visual comparison on the Rain200H dataset for image deraining task. Please zoom in to see the [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Visual comparison on the ablation study. [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Degradation comparison of different IR tasks. upsampling blocks and 1-level wavelet transform in the bottleneck blocks. This setup achieves superior performance, because after downsample layers, the feature sizes are already relatively small, making it challenging for higher levels of DWT to extract meaningful information. Effect of different network architecture [PITH_FULL_IMAGE:figures/full_fig_p018_8.png] view at source ↗
read the original abstract

Spiking neural networks (SNNs) have garnered significant interest in computer vision due to their potential for efficiency and biological inspiration. While spiking CNN-based methods have shown promise for image restoration (IR) tasks, their performance is constrained by the inherent receptive field limitations of CNN operations. In the paper, we explore the benefits of discrete wavelet transformation and propose a spiking pyramid wavelet-based model (SPWM) for high-efficient and low-energy target. Specifically, we develop a spiking dual pyramid wavelet (SDPW) block to model long-range dependency and exploit the properties of the degradation in the wavelet domain. Experimental results on several benchmarks demonstrate that SPWM significantly lowers computational costs and energy consumption while maintaining image quality. Our method showcases the potential of SNNs in the field of IR, offering new insights for future applications of resource-limited devices.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes a spiking pyramid wavelet-based model (SPWM) for image restoration. It develops a spiking dual pyramid wavelet (SDPW) block that combines discrete wavelet transformation with spiking neural networks to model long-range dependencies and exploit degradation properties in the wavelet domain. The central claim, stated in the abstract, is that SPWM significantly lowers computational costs and energy consumption on several benchmarks while maintaining image quality, demonstrating the potential of SNNs for resource-limited IR applications.

Significance. If the experimental claims hold with proper controls, the integration of wavelet-domain processing with spiking mechanisms could offer a meaningful route to low-energy IR models. The work would then contribute to the growing literature on efficient SNN architectures for vision tasks. However, the provided text supplies no quantitative metrics, baselines, energy figures, or ablation results, so the significance cannot be assessed from the manuscript as presented.

major comments (2)
  1. [Abstract] Abstract: the claim that 'experimental results on several benchmarks demonstrate that SPWM significantly lowers computational costs and energy consumption while maintaining image quality' is unsupported; no tables, no energy metrics (e.g., spike counts or mJ/image), no baseline comparisons, no dataset details, and no error bars or ablation studies are supplied anywhere in the text.
  2. [Abstract] Abstract: the assumption that the SDPW block successfully models long-range dependency and exploits wavelet-domain degradation properties without major trade-offs cannot be evaluated, because no equations, architecture diagrams, or implementation details for the SDPW block are given.
minor comments (1)
  1. The title refers to 'Spiking Pyramid Wavelet Transformation' while the abstract uses 'spiking pyramid wavelet-based model (SPWM)'; ensure consistent nomenclature throughout.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed review and for identifying areas where the manuscript requires stronger support for its claims. We address each major comment below and commit to revisions that will incorporate the requested details without altering the core contributions.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that 'experimental results on several benchmarks demonstrate that SPWM significantly lowers computational costs and energy consumption while maintaining image quality' is unsupported; no tables, no energy metrics (e.g., spike counts or mJ/image), no baseline comparisons, no dataset details, and no error bars or ablation studies are supplied anywhere in the text.

    Authors: The referee is correct that the current manuscript text does not contain these supporting elements. We will revise the experimental section to add tables reporting PSNR/SSIM, FLOPs, parameters, spike counts, and energy estimates (mJ/image) on standard benchmarks (e.g., BSD68, Set5, Urban100), with direct comparisons to both SNN and ANN baselines, error bars from multiple runs, and ablation studies on the wavelet and spiking components. This will fully substantiate the abstract claim. revision: yes

  2. Referee: [Abstract] Abstract: the assumption that the SDPW block successfully models long-range dependency and exploits wavelet-domain degradation properties without major trade-offs cannot be evaluated, because no equations, architecture diagrams, or implementation details for the SDPW block are given.

    Authors: We agree that the abstract alone provides insufficient detail for evaluation. The revised manuscript will include the explicit equations governing the discrete wavelet transform integrated with spiking neurons, a clear architecture diagram of the dual-pyramid structure, and implementation specifics (neuron model, threshold, pyramid levels, and how long-range dependencies are captured via multi-scale wavelet coefficients). This will allow readers to assess the modeling of dependencies and degradation properties. revision: yes

Circularity Check

0 steps flagged

No significant circularity; claims rest on external experiments

full rationale

The paper proposes an architectural model (SPWM with SDPW block) combining spiking networks and wavelet transforms for image restoration. Its central claims concern empirical performance on benchmarks (lower costs/energy while preserving quality). No equations, fitted parameters, or derivation steps are shown that reduce by construction to inputs, self-definitions, or self-citations. The load-bearing elements are external experimental comparisons, which are independent of any internal redefinition and therefore do not trigger circularity under the specified criteria.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no equations, training details, or model specifications are provided to identify free parameters, axioms, or invented entities.

pith-pipeline@v0.9.1-grok · 5685 in / 1012 out tokens · 25102 ms · 2026-06-26T21:51:41.770842+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

67 extracted references · 3 linked inside Pith

  1. [1]

    Jiang, B

    H. Jiang, B. Guan, Z. Liu, X. Liu, J. Yu, Z. Liu, S. Han, S. Liu, Learning to see in the extremely dark, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2025, pp. 7676–7685

    2025

  2. [2]

    R. Wang, Y . Zheng, Z. Zhang, C. Li, S. Liu, G. Zhai, X. Liu, Learning hazing to dehazing: Towards realistic haze generation for real-world image dehazing, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025, pp. 23091–23100

    2025

  3. [3]

    X. Liu, Y . Ma, Z. Shi, J. Chen, Griddehazenet: Attention-based multi-scale net- work for image dehazing, in: Proceedings of the IEEE/CVF international confer- ence on computer vision, 2019, pp. 7314–7323

    2019

  4. [4]

    C. Zhao, W. Cai, C. Hu, Z. Yuan, Cycle contrastive adversarial learning with structural consistency for unsupervised high-quality image deraining transformer, Neural Networks (2024) 106428

    2024

  5. [5]

    Y . Tai, R. Xie, C. Zhao, K. Zhang, Z. Zhang, J. Zhou, J. Yang, Addsr: Accelerat- ing diffusion-based blind super-resolution with adversarial diffusion distillation, Pattern Recognition (2026) 113012

    2026

  6. [6]

    Y . Tai, J. Yang, X. Liu, Image super-resolution via deep recursive residual net- work, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 3147–3155

    2017

  7. [7]

    Liang, J

    J. Liang, J. Cao, G. Sun, K. Zhang, L. Van Gool, R. Timofte, Swinir: Image restoration using swin transformer, in: Proceedings of the IEEE/CVF interna- tional conference on computer vision, 2021, pp. 1833–1844

    2021

  8. [8]

    C. Zhao, C. Dong, W. Cai, Y . Wang, Learning a physical-aware diffusion model based on transformer for underwater image enhancement, IEEE Transactions on Geoscience and Remote Sensing 64 (2026) 1–14. 21

    2026

  9. [9]

    Z. Zeng, C. Zhao, W. Cai, Y . Guo, Semantic-guided coarse-to-fine diffusion model for self-supervised image shadow removal, in: International Conference on Computational Visual Media, Springer, 2025, pp. 271–293

    2025

  10. [10]

    Zhao, W.-L

    C. Zhao, W.-L. Cai, Z. Yuan, C.-W. Hu, Multi-cropping contrastive learning and domain consistency for unsupervised image-to-image translation, IET Image Pro- cessing 19 (1) (2025) e70006

    2025

  11. [11]

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, M.-H. Yang, L. Shao, Multi-stage progressive image restoration, in: Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, 2021, pp. 14821–14831

    2021

  12. [12]

    X. Chen, H. Li, M. Li, J. Pan, Learning a sparse transformer network for effec- tive image deraining, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 5896–5905

    2023

  13. [13]

    M. Yao, J. Hu, T. Hu, Y . Xu, Z. Zhou, Y . Tian, B. XU, G. Li, Spike-driven transformer v2: Meta spiking neural network architecture inspiring the design of next-generation neuromorphic chips, in: The Twelfth International Conference on Learning Representations, 2024

    2024

  14. [14]

    Q. Su, Y . Chou, Y . Hu, J. Li, S. Mei, Z. Zhang, G. Li, Deep directly-trained spiking neural networks for object detection, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 6555–6565

    2023

  15. [15]

    K. Roy, A. Jaiswal, P. Panda, Towards spike-based machine intelligence with neuromorphic computing, Nature 575 (7784) (2019) 607–617

    2019

  16. [16]

    T. Song, G. Jin, P. Li, K. Jiang, X. Chen, J. Jin, Learning a spiking neural network for efficient image deraining, in: Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, 2024, pp. 1254–1262

    2024

  17. [17]

    J. Xiao, X. Fu, Y . Zhu, D. Li, J. Huang, K. Zhu, Z.-J. Zha, Homoformer: Homog- enized transformer for image shadow removal, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 25617– 25626. 22

    2024

  18. [18]

    W. Luo, Y . Li, R. Urtasun, R. Zemel, Understanding the effective receptive field in deep convolutional neural networks, Advances in neural information processing systems 29 (2016)

    2016

  19. [19]

    M. Yao, G. Zhao, H. Zhang, Y . Hu, L. Deng, Y . Tian, B. Xu, G. Li, Attention spiking neural networks, IEEE transactions on pattern analysis and machine in- telligence 45 (8) (2023) 9393–9410

    2023

  20. [20]

    W. Dong, H. Zhou, Y . Zhang, X. Liu, J. Chen, Ecmamba: Consolidating selective state space model with retinex guidance for efficient multiple exposure correction, Advances in Neural Information Processing Systems 37 (2024) 53438–53457

    2024

  21. [21]

    C. Zhao, W. Cai, C. Dong, Z. Zeng, Toward sufficient spatial-frequency inter- action for gradient-aware underwater image enhancement, in: ICASSP 2024- 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, 2024, pp. 3220–3224

    2024

  22. [22]

    H. Zhou, W. Dong, X. Liu, S. Liu, X. Min, G. Zhai, J. Chen, Glare: Low light image enhancement via generative latent feature based codebook retrieval, in: European Conference on Computer Vision, Springer, 2024, pp. 36–54

    2024

  23. [23]

    H. Zhou, W. Dong, X. Liu, Y . Zhang, G. Zhai, J. Chen, Low-light image enhance- ment via generative perceptual priors, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 39, 2025, pp. 10752–10760

    2025

  24. [24]

    Zhao, W.-L

    C. Zhao, W.-L. Cai, Z. Yuan, Spectral normalization and dual contrastive regular- ization for image-to-image translation, The Visual Computer 41 (2025) 129–140

    2025

  25. [25]

    S. Lu, Z. Lian, Z. Zhou, S. Zhang, C. Zhao, A. W.-K. Kong, Does flux already know how to perform physically plausible image composition?, arXiv preprint arXiv:2509.21278 (2025)

    arXiv 2025

  26. [26]

    C. Dong, C. Zhao, W. Cai, B. Yang, Y . Guo, O-mamba: O-shape state-space model for underwater image enhancement, in: Chinese Conference on Pattern Recognition and Computer Vision (PRCV), Springer, 2025, pp. 168–182. 23

    2025

  27. [27]

    X. Zhao, C. Zhao, X. Hu, H. Zhang, Y . Tai, J. Yang, Learning multi-scale spatial- frequency features for image denoising, Pattern Recognition (2025) 112300

    2025

  28. [28]

    C. Dong, C. C. Loy, K. He, X. Tang, Image super-resolution using deep convolu- tional networks, IEEE transactions on pattern analysis and machine intelligence 38 (2) (2015) 295–307

    2015

  29. [29]

    Zhang, W

    K. Zhang, W. Zuo, Y . Chen, D. Meng, L. Zhang, Beyond a gaussian denoiser: Residual learning of deep cnn for image denoising, IEEE transactions on image processing 26 (7) (2017) 3142–3155

    2017

  30. [30]

    D. Ren, W. Zuo, Q. Hu, P. Zhu, D. Meng, Progressive image deraining networks: A better and simpler baseline, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 3937–3946

    2019

  31. [31]

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, M.-H. Yang, Restormer: Efficient transformer for high-resolution image restoration, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 5728–5739

    2022

  32. [32]

    Z. Wang, Y . Fang, J. Cao, Q. Zhang, Z. Wang, R. Xu, Masked spiking transformer, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 1761–1771

    2023

  33. [33]

    Z. Zhou, Y . Zhu, C. He, Y . Wang, Y . Shuicheng, Y . Tian, L. Yuan, Spikformer: When spiking neural network meets transformer, in: The Eleventh International Conference on Learning Representations, 2023

    2023

  34. [34]

    C. Zhao, Z. Chen, Y . Xu, E. Gu, J. Li, Z. Yi, Q. Wang, J. Yang, Y . Tai, From zero to detail: Deconstructing ultra-high-definition image restoration from progressive spectral perspective, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025, pp. 17935–17946

    2025

  35. [35]

    C. Zhao, W. Cai, C. Dong, C. Hu, Wavelet-based fourier information interaction with frequency diffusion adjustment for underwater image restoration, in: Pro- 24 ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, 2024, pp. 8281–8291

    2024

  36. [36]

    C. Zhao, J. Chen, H. Li, Z. Kang, S. Lu, X. Wei, K. Zhang, J. Yang, Y . Tai, Luve: Latent-cascaded ultra-high-resolution video generation with dual frequency ex- perts, arXiv preprint arXiv:2602.11564 (2026)

    Pith/arXiv arXiv 2026

  37. [37]

    Jiang, A

    H. Jiang, A. Luo, H. Fan, S. Han, S. Liu, Low-light image enhancement with wavelet-based diffusion models, ACM Transactions on Graphics (TOG) 42 (6) (2023) 1–14

    2023

  38. [38]

    C. Zhao, Y . Xu, Z. Chen, E. Gu, K. Zhang, X. Liu, J. Yang, Y . Tai, From zero to detail: A progressive spectral decoupling paradigm for uhd image restoration with new benchmark, IEEE Transactions on Pattern Analysis and Machine Intelligence (2026) 1–18

    2026

  39. [39]

    C. Zhao, E. Ci, Y . Xu, T. Fan, S. Guan, Y . Ge, J. Yang, Y . Tai, Ultrahr-100k: Enhancing uhr image synthesis with a large-scale high-quality dataset, Advances in Neural Information Processing Systems 38 (2026) 3373–3393

    2026

  40. [40]

    Y . Zhou, C. Zhao, D. Peng, F. Ji, X.-T. Yuan, Learning to refine: Spectral- decoupled iterative refinement framework for precipitation nowcasting, arXiv preprint arXiv:2606.02661 (2026)

    Pith/arXiv arXiv 2026

  41. [41]

    W. Fang, Y . Chen, J. Ding, Z. Yu, T. Masquelier, D. Chen, L. Huang, H. Zhou, G. Li, Y . Tian, Spikingjelly: An open-source machine learning infrastructure plat- form for spike-based intelligence, Science Advances 9 (40) (2023) eadi1480

    2023

  42. [42]

    H. Wang, Q. Xie, Q. Zhao, D. Meng, A model-driven deep neural network for single image rain removal, in: CVPR, 2020, pp. 3103–3112

    2020

  43. [43]

    Q. Guo, J. Sun, F. Juefei-Xu, L. Ma, X. Xie, W. Feng, Y . Liu, J. Zhao, Efficient- derain: Learning pixel-wise dilation filtering for high-efficiency single-image de- raining, in: AAAI, V ol. 35, 2021, pp. 1487–1495. 25

    2021

  44. [44]

    Y . Cui, Y . Tao, Z. Bing, W. Ren, X. Gao, X. Cao, K. Huang, A. Knoll, Selective frequency network for image restoration, in: ICLR, 2023

    2023

  45. [45]

    Liang, S

    Y . Liang, S. Anwar, Y . Liu, Drt: A lightweight single image deraining recursive transformer, in: CVPRw, 2022, pp. 589–598

    2022

  46. [46]

    Jiang, Z

    K. Jiang, Z. Wang, C. Chen, Z. Wang, L. Cui, C.-W. Lin, Magic elf: Image deraining meets association learning and transformer, in: ACM MM, 2022, pp. 827–836

    2022

  47. [47]

    Q. Wang, K. Jiang, J. Lai, Z. Wang, J. Zhang, Hpcnet: A hybrid progressive coupled network for image deraining, in: 2023 IEEE International Conference on Multimedia and Expo (ICME), IEEE, 2023, pp. 2747–2752

    2023

  48. [48]

    Y . Wang, C. Ma, J. Liu, Smartassign: Learning a smart knowledge assignment strategy for deraining and desnowing, in: CVPR, 2023, pp. 3677–3686

    2023

  49. [49]

    C. Wei, W. Wang, W. Yang, J. Liu, Deep retinex decomposition for low-light enhancement, arXiv preprint arXiv:1808.04560 (2018)

    Pith/arXiv arXiv 2018

  50. [50]

    W. Yang, S. Wang, Y . Fang, Y . Wang, J. Liu, From fidelity to perceptual quality: A semi-supervised approach for low-light image enhancement, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 3063–3072

    2020

  51. [51]

    Jiang, X

    Y . Jiang, X. Gong, D. Liu, Y . Cheng, C. Fang, X. Shen, J. Yang, P. Zhou, Z. Wang, Enlightengan: Deep light enhancement without paired supervision, IEEE trans- actions on image processing 30 (2021) 2340–2349

    2021

  52. [52]

    W. Wang, C. Wei, W. Yang, J. Liu, Gladnet: Low-light enhancement network with global awareness, in: 2018 13th IEEE international conference on automatic face & gesture recognition (FG 2018), IEEE, 2018, pp. 751–755

    2018

  53. [53]

    Zhang, J

    Y . Zhang, J. Zhang, X. Guo, Kindling the darkness: A practical low-light image enhancer, in: Proceedings of the 27th ACM international conference on multime- dia, 2019, pp. 1632–1640. 26

    2019

  54. [54]

    Zhang, X

    Y . Zhang, X. Guo, J. Ma, W. Liu, J. Zhang, Beyond brightening low-light images, International Journal of Computer Vision 129 (2021) 1013–1037

    2021

  55. [55]

    Z. Wang, X. Cun, J. Bao, W. Zhou, J. Liu, H. Li, Uformer: A general u-shaped transformer for image restoration, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 17683–17693

    2022

  56. [56]

    T. Wang, K. Zhang, T. Shen, W. Luo, B. Stenger, T. Lu, Ultra-high-definition low-light image enhancement: A benchmark and transformer-based method, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 37, 2023, pp. 2654–2662

    2023

  57. [57]

    Shakibania, S

    H. Shakibania, S. Raoufi, H. Khotanlou, Cdan: convolutional dense attention- guided network for low-light image enhancement, Digital Signal Processing 156 (2025) 104802

    2025

  58. [58]

    Y . Gou, P. Hu, J. Lv, J. T. Zhou, X. Peng, Multi-scale adaptive network for single image denoising, Advances in Neural Information Processing Systems 35 (2022) 14099–14112

    2022

  59. [59]

    C. Ren, X. He, C. Wang, Z. Zhao, Adaptive consistency prior based deep network for image denoising, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 8596–8606

    2021

  60. [60]

    W. Wu, S. Liu, Y . Xia, Y . Zhang, Dual residual attention network for image de- noising, Pattern Recognition 149 (2024) 110291

    2024

  61. [61]

    Zhang, Y

    J. Zhang, Y . Zhang, J. Gu, J. Dong, L. Kong, X. Yang, Xformer: Hybrid x-shaped transformer for image denoising, in: The Twelfth International Conference on Learning Representations, 2024

    2024

  62. [62]

    W. Yang, R. T. Tan, J. Feng, J. Liu, Z. Guo, S. Yan, Deep joint rain detection and removal from a single image, in: CVPR, 2017, pp. 1357–1366

    2017

  63. [63]

    Zhang, V

    H. Zhang, V . M. Patel, Density-aware single image de-raining using a multi- stream dense network, in: CVPR, 2018, pp. 695–704. 27

    2018

  64. [64]

    C. O. Ancuti, C. Ancuti, M. Sbert, R. Timofte, Dense-haze: A benchmark for im- age dehazing with dense-haze and haze-free images, in: 2019 IEEE international conference on image processing (ICIP), IEEE, 2019, pp. 1014–1018

    2019

  65. [65]

    X. Qin, Z. Wang, Y . Bai, X. Xie, H. Jia, Ffa-net: Feature fusion attention network for single image dehazing, in: Proceedings of the AAAI conference on artificial intelligence, V ol. 34, 2020, pp. 11908–11915

    2020

  66. [66]

    H. Dong, J. Pan, L. Xiang, Z. Hu, X. Zhang, F. Wang, M.-H. Yang, Multi- scale boosted dehazing network with dense feature fusion, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 2157–2167

    2020

  67. [67]

    M. Zhou, J. Huang, C.-L. Guo, C. Li, Fourmer: An efficient global modeling paradigm for image restoration, in: International conference on machine learning, PMLR, 2023, pp. 42589–42601. 28

    2023