pith. sign in

arxiv: 2605.13158 · v1 · pith:FCRWSQGQnew · submitted 2026-05-13 · 💻 cs.CV

Unifying Physically-Informed Weather Priors in A Single Model for Image Restoration Across Multiple Adverse Weather Conditions

Jiaqi Xu , Xiaowei Hu , Lei Zhu , Pheng-Ann Heng This is my paper

Pith reviewed 2026-05-14 19:10 UTC · model grok-4.3

classification 💻 cs.CV
keywords image restorationadverse weatherunified modelweather priorsocclusion maptransmission mapdeep learning
0
0 comments X

The pith

A unified imaging model accounting for both visible particles and aggregate fog scattering restores images across multiple adverse weather conditions in one network.

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

The paper seeks to replace multiple specialized networks with one model that recovers clear scenes from rain, snow, fog, and similar degradations. It identifies shared physical factors: close particles act as individual occlusions while distant ones create fog-like scattering. By deriving occlusion and transmission maps from a single unified imaging model, the network can enhance features without weather-specific retraining. A reader would care because outdoor vision systems in vehicles or surveillance often encounter mixed or changing conditions that defeat current per-weather solutions.

Core claim

We analyze the common visual factors in adverse weather conditions and present a unified imaging model that considers the individually visible particles and fog-like aggregate scattering effects. Further, we design a novel weather-prior-based network, which leverages the weather-related prior information to help recover the scene by enhancing the features using the estimated occlusion and transmission.

What carries the argument

Unified imaging model that produces occlusion and transmission maps from shared particle and aggregate scattering effects to guide feature enhancement in a single network.

If this is right

  • The single network outperforms prior state-of-the-art methods on standard benchmarks covering multiple adverse weather scenarios.
  • Restoration succeeds across weather types because the same occlusion and transmission estimates drive feature recovery without per-condition adjustments.
  • Weather-related prior information extracted once from the unified model is reused to enhance degraded features in every tested condition.

Where Pith is reading between the lines

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

  • The same distance-dependent occlusion and scattering logic could be tested on dust or smoke degradations that follow similar physical rules.
  • The occlusion and transmission estimates might serve as auxiliary inputs to improve downstream tasks such as object detection in adverse weather.
  • If the priors remain stable under small camera or lighting changes, the model could support video restoration without frame-by-frame retuning.

Load-bearing premise

That a single set of estimated occlusion and transmission maps derived from the unified model will be sufficient and accurate enough to guide feature enhancement across all tested weather types without condition-specific tuning or post-hoc adjustments.

What would settle it

Apply the trained model to images containing a previously unseen weather mixture, such as dense mist with falling particles, and check whether PSNR and SSIM fall below those of separately trained specialized models.

Figures

Figures reproduced from arXiv: 2605.13158 by Jiaqi Xu, Lei Zhu, Pheng-Ann Heng, Xiaowei Hu.

Figure 1
Figure 1. Figure 1: Examples of real photos for (a) haze, (b) rain, and [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The visual process in adverse weather conditions. The [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The overall architecture of our WeatherNet. (a) WeatherNet contains a weather-related prior estimation U-Net [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The volumetric effects α of rain and snow are con￾sidered for the near regions, αnear, and far regions, αfar, to synthesize the degraded image I, where αnear is computed with multiple layers based on depth d. Overall, the rain and snow imaging models follow a similar footprint in the recent deep learning-based studies, i.e., with only rain streaks or snowflakes in Eqs. (2) and (4), or with scattering effec… view at source ↗
Figure 5
Figure 5. Figure 5: Transmission-guided global attention (TGGA) and [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Samples of our Weather30K dataset, shown with [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Visual comparisons on the synthetic datasets. The first three rows are from our Weather30K dataset, while the last row is from the Rain-Haze dataset [8]. Please zoom in for a better view. TABLE I: Quantitative comparisons on our Weather30K (the first experimental setting). The best and second-best results are marked using bold and underline, respectively. Type Method PSNR ↑ / SSIM ↑ Haze Rain Snow Mixed Tr… view at source ↗
Figure 8
Figure 8. Figure 8: Visual comparisons of real-world photos under adverse weather conditions. Our method clearly removes haze, rain, and snow artifacts and generates more visually appealing results. Please zoom in for a better view. TABLE IV: No-reference evaluation results on the real-world RTTS [3], DDN-SIRR [87], and Snow100K [9] datasets. Method NIQE ↓ / MUSIQ ↑ Haze Rain Snow TransWeather [13] 6.29/47.17 3.88/55.24 3.56/… view at source ↗
Figure 11
Figure 11. Figure 11: Ablation study results of the proposed framework. The weather-related artifacts of inputs are greatly reduced in the initially restored images, and our two-stage network further refines the clear scenes. TABLE VI: Ablation study on the overall framework and the major components design. (a) (b) (c) (d) (e) Our method U-Net θe ✓ ✓ ✓ ✓ ✓ U-Net θr ✓ ✓ ✓ ✓ ✓ TGGA ✓ ✓ OGLA ✓ ✓ PSNR 24.86 23.97 25.63 26.70 26.47… view at source ↗
Figure 10
Figure 10. Figure 10: Ablation study results of the imaging model. The model trained with data synthesized by our imaging model, considering the rain and snow volumetric effects with multiple layers, removes more visible particles in real-world images. visual factors like occlusions and scattering effects in haze, rain, and snow, while incorporating the volumetric effects of rain and snow particles. First, as shown in [PITH_F… view at source ↗
Figure 12
Figure 12. Figure 12: Ablation on the guidance information. Our method with transmission and occlusion guidance improves the visual quality of restoration results for ambiguous regions. TABLE VII: Ablation study on the guidance information in our WACA module. Method PSNR Haze Rain Snow Basic 26.65 28.02 25.59 Basic+Transmission 26.92 28.33 25.86 Basic+Occlusion 26.79 28.41 25.93 Our Method 27.01 28.56 26.09 either transmission… view at source ↗
Figure 13
Figure 13. Figure 13: Visualization of downstream application results be [PITH_FULL_IMAGE:figures/full_fig_p011_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: GPT-4o is used to describe and compare the input real-world images (first row) and our restoration results (second [PITH_FULL_IMAGE:figures/full_fig_p012_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: A failure case with extreme adverse weather effects. [PITH_FULL_IMAGE:figures/full_fig_p012_15.png] view at source ↗
read the original abstract

Image restoration under multiple adverse weather conditions aims to develop a single model to recover the underlying scene with high visibility. Weather-related artifacts vary with the particle's distance to the camera according to the established scene visibility analysis, where close and faraway regions are more affected by falling drops and fog effects, respectively. Existing methods fail to consider this weather-specific physical visual process; thus, the restoration performance is limited. In this work, we analyze the common visual factors in adverse weather conditions and present a unified imaging model that considers the individually visible particles and fog-like aggregate scattering effects. Further, we design a novel weather-prior-based network, which leverages the weather-related prior information to help recover the scene by enhancing the features using the estimated occlusion and transmission. Experimental results in multiple adverse scenarios show the superiority of our method against state-of-the-art methods.

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 unified physically-informed imaging model for restoring images degraded by multiple adverse weather conditions (e.g., rain, fog). It identifies common visual factors based on established scene-visibility analysis—individually visible particles affecting close regions and fog-like aggregate scattering affecting distant regions—and introduces a weather-prior-based network that estimates occlusion and transmission maps to guide feature enhancement and scene recovery. The abstract claims experimental superiority over state-of-the-art methods across adverse scenarios.

Significance. If the quantitative results and ablations hold, the work offers a meaningful step toward single-model generalization across weather degradations rather than condition-specific pipelines. Grounding the priors in scene-visibility physics is a strength that could improve robustness in real-world vision systems such as autonomous driving and outdoor surveillance. The approach avoids re-deriving fitted quantities from scratch and instead leverages established analysis, which is a positive design choice.

major comments (2)
  1. [Abstract] Abstract: The assertion of superiority is unsupported by any quantitative metrics, error bars, ablation results, or dataset details. This makes the central claim impossible to verify from the provided text and requires the full experimental section (including tables of PSNR/SSIM, cross-weather comparisons, and ablation on the occlusion/transmission maps) to be load-bearing for acceptance.
  2. [Abstract] The manuscript relies on the assumption that a single set of estimated occlusion and transmission maps derived from the unified model will be accurate and sufficient across all weather types without condition-specific tuning. No evidence or ablation is referenced in the abstract to support this; if the full text lacks such controls, the generalization claim is at risk.
minor comments (1)
  1. Clarify the exact mathematical form of the unified imaging model (e.g., how particle and aggregate scattering terms are combined) and ensure all symbols are defined on first use.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address the two major comments on the abstract below and will revise the manuscript to strengthen the presentation of results and evidence.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The assertion of superiority is unsupported by any quantitative metrics, error bars, ablation results, or dataset details. This makes the central claim impossible to verify from the provided text and requires the full experimental section (including tables of PSNR/SSIM, cross-weather comparisons, and ablation on the occlusion/transmission maps) to be load-bearing for acceptance.

    Authors: We agree that the abstract does not contain specific quantitative metrics. The full manuscript (Section 4) contains the required experimental details, including PSNR/SSIM tables across multiple weather conditions and datasets, cross-weather comparisons, and ablations on the occlusion and transmission maps. To address the concern, we will revise the abstract to include a concise statement of key quantitative gains (e.g., average PSNR improvement) while remaining within length limits, thereby making the superiority claim more verifiable from the abstract itself. revision: yes

  2. Referee: [Abstract] The manuscript relies on the assumption that a single set of estimated occlusion and transmission maps derived from the unified model will be accurate and sufficient across all weather types without condition-specific tuning. No evidence or ablation is referenced in the abstract to support this; if the full text lacks such controls, the generalization claim is at risk.

    Authors: The unified imaging model in Section 3 is derived from common physical factors (individually visible particles and aggregate scattering) that hold across weather types per scene-visibility analysis, and the weather-prior network estimates a single set of occlusion and transmission maps without per-condition tuning. The full manuscript provides supporting ablations (Section 4.3) showing these maps remain effective across rain, fog, and other conditions. We will revise the abstract to briefly reference these generalization results from the experiments. revision: yes

Circularity Check

0 steps flagged

Unified model grounded in established scene-visibility priors; no circular reduction

full rationale

The derivation begins from the established scene-visibility analysis (close regions affected by drops, distant by fog) and introduces a unified imaging model that explicitly separates individually visible particles from aggregate scattering. Occlusion and transmission maps are estimated from this model to guide feature enhancement, but the paper does not fit parameters to a subset of outputs and then relabel them as predictions, nor does it rely on self-citation chains for uniqueness or ansatz. The central claim therefore retains independent physical content and does not reduce to its own inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach rests on the domain assumption of distance-dependent visibility effects and introduces no new free parameters or invented entities beyond standard network weights.

axioms (1)
  • domain assumption Scene visibility varies with particle distance to the camera, with close regions dominated by falling drops and far regions by fog-like scattering
    Invoked in the abstract to motivate the unified imaging model

pith-pipeline@v0.9.0 · 5454 in / 1060 out tokens · 46324 ms · 2026-05-14T19:10:28.438887+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

95 extracted references · 95 canonical work pages · 1 internal anchor

  1. [1]

    Single image haze removal using dark channel prior,

    K. He, J. Sun, and X. Tang, “Single image haze removal using dark channel prior,”TPAMI, 2010

    work page 2010

  2. [2]

    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,”TIP, 2016

    work page 2016

  3. [3]

    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,”TIP, 2018

    work page 2018

  4. [4]

    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,” inCVPR, 2021

    work page 2021

  5. [5]

    Vision and rain,

    K. Garg and S. K. Nayar, “Vision and rain,”IJCV, 2007

    work page 2007

  6. [6]

    Removing rain from single images via a deep detail network,

    X. Fu, J. Huang, D. Zeng, Y . Huang, X. Ding, and J. Paisley, “Removing rain from single images via a deep detail network,” inCVPR, 2017

    work page 2017

  7. [7]

    Attentive generative adversarial network for raindrop removal from a single image,

    R. Qian, R. T. Tan, W. Yang, J. Su, and J. Liu, “Attentive generative adversarial network for raindrop removal from a single image,” inCVPR, 2018

    work page 2018

  8. [8]

    Heavy rain image restoration: Integrating physics model and conditional adversarial learning,

    R. Li, L.-F. Cheong, and R. T. Tan, “Heavy rain image restoration: Integrating physics model and conditional adversarial learning,” in CVPR, 2019

    work page 2019

  9. [9]

    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,”TIP, 2018

    work page 2018

  10. [10]

    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,” inECCV, 2020

    work page 2020

  11. [11]

    Deep dense multi-scale network for snow removal using semantic and depth priors,

    K. Zhang, R. Li, Y . Yu, W. Luo, and C. Li, “Deep dense multi-scale network for snow removal using semantic and depth priors,”TIP, 2021

    work page 2021

  12. [12]

    All in one bad weather removal using architectural search,

    R. Li, R. T. Tan, and L.-F. Cheong, “All in one bad weather removal using architectural search,” inCVPR, 2020

    work page 2020

  13. [13]

    Transweather: Transformer-based restoration of images degraded by adverse weather conditions,

    J. M. J. Valanarasu, R. Yasarla, and V . M. Patel, “Transweather: Transformer-based restoration of images degraded by adverse weather conditions,” inCVPR, 2022

    work page 2022

  14. [14]

    Learning multiple adverse weather removal via two-stage knowl- edge learning and multi-contrastive regularization: Toward a unified model,

    W.-T. Chen, Z.-K. Huang, C.-C. Tsai, H.-H. Yang, J.-J. Ding, and S.-Y . Kuo, “Learning multiple adverse weather removal via two-stage knowl- edge learning and multi-contrastive regularization: Toward a unified model,” inCVPR, 2022

    work page 2022

  15. [15]

    Restoring vision in adverse weather conditions with patch-based denoising diffusion models,

    O. ¨Ozdenizci and R. Legenstein, “Restoring vision in adverse weather conditions with patch-based denoising diffusion models,”TPAMI, 2023

    work page 2023

  16. [16]

    Learning weather-general and weather-specific features for image restoration under multiple adverse weather conditions,

    Y . Zhu, T. Wang, X. Fu, X. Yang, X. Guo, J. Dai, Y . Qiao, and X. Hu, “Learning weather-general and weather-specific features for image restoration under multiple adverse weather conditions,” inCVPR, 2023

    work page 2023

  17. [17]

    Depth-attentional features for single-image rain removal,

    X. Hu, C.-W. Fu, L. Zhu, and P.-A. Heng, “Depth-attentional features for single-image rain removal,” inCVPR, 2019

    work page 2019

  18. [18]

    Vision and the atmosphere,

    S. G. Narasimhan and S. K. Nayar, “Vision and the atmosphere,”IJCV, 2002

    work page 2002

  19. [19]

    Densely connected pyramid dehazing network,

    H. Zhang and V . M. Patel, “Densely connected pyramid dehazing network,” inCVPR, 2018

    work page 2018

  20. [20]

    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,” inAAAI, 2020

    work page 2020

  21. [21]

    Multi-scale boosted dehazing network with dense feature fusion,

    H. Dong, J. Pan, L. Xiang, Z. Hu, X. Zhang, F. Wang, and M.-H. Yang, “Multi-scale boosted dehazing network with dense feature fusion,” in CVPR, 2020

    work page 2020

  22. [22]

    Vision transformers for single image dehazing,

    Y . Song, Z. He, H. Qian, and X. Du, “Vision transformers for single image dehazing,”TIP, 2023

    work page 2023

  23. [23]

    Single image dehazing based on the physical model and msrcr algorithm,

    J. Wang, K. Lu, J. Xue, N. He, and L. Shao, “Single image dehazing based on the physical model and msrcr algorithm,”Transactions on Circuits and Systems for Video Technology, 2017

    work page 2017

  24. [24]

    A variational framework for underwater image dehazing and deblurring,

    J. Xie, G. Hou, G. Wang, and Z. Pan, “A variational framework for underwater image dehazing and deblurring,”Transactions on Circuits and Systems for Video Technology, 2021

    work page 2021

  25. [25]

    Deep joint rain detection and removal from a single image,

    W. Yang, R. T. Tan, J. Feng, J. Liu, Z. Guo, and S. Yan, “Deep joint rain detection and removal from a single image,” inCVPR, 2017

    work page 2017

  26. [26]

    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,” inECCV, 2018

    work page 2018

  27. [27]

    Spatial attentive single-image deraining with a high quality real rain dataset,

    T. Wang, X. Yang, K. Xu, S. Chen, Q. Zhang, and R. W. Lau, “Spatial attentive single-image deraining with a high quality real rain dataset,” inCVPR, 2019

    work page 2019

  28. [28]

    Multi-scale progressive fusion network for single image deraining,

    K. Jiang, Z. Wang, P. Yi, C. Chen, B. Huang, Y . Luo, J. Ma, and J. Jiang, “Multi-scale progressive fusion network for single image deraining,” in CVPR, 2020

    work page 2020

  29. [29]

    Single image deraining: From model-based to data-driven and beyond,

    W. Yang, R. T. Tan, S. Wang, Y . Fang, and J. Liu, “Single image deraining: From model-based to data-driven and beyond,”TPAMI, 2020

    work page 2020

  30. [30]

    Adherent raindrop modeling, detection and removal in video,

    S. You, R. T. Tan, R. Kawakami, Y . Mukaigawa, and K. Ikeuchi, “Adherent raindrop modeling, detection and removal in video,”TPAMI, 2015

    work page 2015

  31. [31]

    Single-image real- time rain removal based on depth-guided non-local features,

    X. Hu, L. Zhu, T. Wang, C.-W. Fu, and P.-A. Heng, “Single-image real- time rain removal based on depth-guided non-local features,”TIP, 2021

    work page 2021

  32. [32]

    Decomposition makes better rain removal: An improved attention- guided deraining network,

    K. Jiang, Z. Wang, P. Yi, C. Chen, Z. Han, T. Lu, B. Huang, and J. Jiang, “Decomposition makes better rain removal: An improved attention- guided deraining network,”Transactions on Circuits and Systems for Video Technology, 2020

    work page 2020

  33. [33]

    Subjective and objective de-raining quality assessment towards authentic rain image,

    Q. Wu, L. Wang, K. N. Ngan, H. Li, F. Meng, and L. Xu, “Subjective and objective de-raining quality assessment towards authentic rain image,” Transactions on Circuits and Systems for Video Technology, 2020

    work page 2020

  34. [34]

    Analysis of rain and snow in frequency space,

    P. C. Barnum, S. Narasimhan, and T. Kanade, “Analysis of rain and snow in frequency space,”IJCV, 2010

    work page 2010

  35. [35]

    All snow removed: Single image desnowing algo- rithm using hierarchical dual-tree complex wavelet representation and contradict channel loss,

    W.-T. Chen, H.-Y . Fang, C.-L. Hsieh, C.-C. Tsai, I. Chen, J.-J. Ding, S.-Y . Kuoet al., “All snow removed: Single image desnowing algo- rithm using hierarchical dual-tree complex wavelet representation and contradict channel loss,” inICCV, 2021

    work page 2021

  36. [36]

    Image desnowing via deep invertible separation,

    Y . Quan, X. Tan, Y . Huang, Y . Xu, and H. Ji, “Image desnowing via deep invertible separation,”TCSVT, 2023

    work page 2023

  37. [37]

    Single image dehazing,

    R. Fattal, “Single image dehazing,”TOG, 2008

    work page 2008

  38. [38]

    Non-local image dehazing,

    D. Berman, S. Avidanet al., “Non-local image dehazing,” inCVPR, 2016

    work page 2016

  39. [39]

    Single image dehazing via multi-scale convolutional neural networks,

    W. Ren, S. Liu, H. Zhang, J. Pan, X. Cao, and M.-H. Yang, “Single image dehazing via multi-scale convolutional neural networks,” in ECCV, 2016

    work page 2016

  40. [40]

    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,” inICCV, 2017

    work page 2017

  41. [41]

    Single image dehazing via conditional generative adversarial network,

    R. Li, J. Pan, Z. Li, and J. Tang, “Single image dehazing via conditional generative adversarial network,” inCVPR, 2018

    work page 2018

  42. [42]

    Enhanced pix2pix dehazing network,

    Y . Qu, Y . Chen, J. Huang, and Y . Xie, “Enhanced pix2pix dehazing network,” inCVPR, 2019

    work page 2019

  43. [43]

    Gated fusion network for single image dehazing,

    W. Ren, L. Ma, J. Zhang, J. Pan, X. Cao, W. Liu, and M.-H. Yang, “Gated fusion network for single image dehazing,” inCVPR, 2018

    work page 2018

  44. [44]

    Deep multi-model fusion for single-image dehazing,

    Z. Deng, L. Zhu, X. Hu, C.-W. Fu, X. Xu, Q. Zhang, J. Qin, and P.-A. Heng, “Deep multi-model fusion for single-image dehazing,” inICCV, 2019

    work page 2019

  45. [45]

    Griddehazenet: Attention-based multi-scale network for image dehazing,

    X. Liu, Y . Ma, Z. Shi, and J. Chen, “Griddehazenet: Attention-based multi-scale network for image dehazing,” inICCV, 2019

    work page 2019

  46. [46]

    Gated context aggregation network for image dehazing and deraining,

    D. Chen, M. He, Q. Fan, J. Liao, L. Zhang, D. Hou, L. Yuan, and G. Hua, “Gated context aggregation network for image dehazing and deraining,” inWACV, 2019

    work page 2019

  47. [47]

    Image dehazing transformer with transmission-aware 3d position embedding,

    C.-L. Guo, Q. Yan, S. Anwar, R. Cong, W. Ren, and C. Li, “Image dehazing transformer with transmission-aware 3d position embedding,” inCVPR, 2022

    work page 2022

  48. [48]

    Perceiving and modeling density for image dehazing,

    T. Ye, Y . Zhang, M. Jiang, L. Chen, Y . Liu, S. Chen, and E. Chen, “Perceiving and modeling density for image dehazing,” inECCV, 2022

    work page 2022

  49. [49]

    Semantic foggy scene under- standing with synthetic data,

    C. Sakaridis, D. Dai, and L. Van Gool, “Semantic foggy scene under- standing with synthetic data,”IJCV, 2018

    work page 2018

  50. [50]

    Distilling image dehazing with heterogeneous task imitation,

    M. Hong, Y . Xie, C. Li, and Y . Qu, “Distilling image dehazing with heterogeneous task imitation,” inCVPR, 2020

    work page 2020

  51. [51]

    Domain adaptation for image dehazing,

    Y . Shao, L. Li, W. Ren, C. Gao, and N. Sang, “Domain adaptation for image dehazing,” inCVPR, 2020

    work page 2020

  52. [52]

    Ultra- high-definition image dehazing via multi-guided bilateral learning,

    Z. Zheng, W. Ren, X. Cao, X. Hu, T. Wang, F. Song, and X. Jia, “Ultra- high-definition image dehazing via multi-guided bilateral learning,” in CVPR, 2021

    work page 2021

  53. [53]

    Psd: Principled synthetic-to- real dehazing guided by physical priors,

    Z. Chen, Y . Wang, Y . Yang, and D. Liu, “Psd: Principled synthetic-to- real dehazing guided by physical priors,” inCVPR, 2021

    work page 2021

  54. [54]

    From synthetic to real: Image dehazing collaborating with unlabeled real data,

    Y . Liu, L. Zhu, S. Pei, H. Fu, J. Qin, Q. Zhang, L. Wan, and W. Feng, “From synthetic to real: Image dehazing collaborating with unlabeled real data,” inACMMM, 2021

    work page 2021

  55. [55]

    Self- augmented unpaired image dehazing via density and depth decompo- sition,

    Y . Yang, C. Wang, R. Liu, L. Zhang, X. Guo, and D. Tao, “Self- augmented unpaired image dehazing via density and depth decompo- sition,” inCVPR, 2022

    work page 2022

  56. [56]

    Utilizing local phase information to remove rain from video,

    V . Santhaseelan and V . K. Asari, “Utilizing local phase information to remove rain from video,”IJCV, 2015

    work page 2015

  57. [57]

    Image de-raining using a conditional generative adversarial network,

    H. Zhang, V . Sindagi, and V . M. Patel, “Image de-raining using a conditional generative adversarial network,”TCSVT, 2019. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 14

    work page 2019

  58. [58]

    Image de-raining transformer,

    J. Xiao, X. Fu, A. Liu, F. Wu, and Z.-J. Zha, “Image de-raining transformer,”TPAMI, 2022

    work page 2022

  59. [59]

    Rethinking image deraining via rain streaks and vapors,

    Y . Wang, Y . Song, C. Ma, and B. Zeng, “Rethinking image deraining via rain streaks and vapors,” inECCV, 2020

    work page 2020

  60. [60]

    From rain generation to rain removal,

    H. Wang, Z. Yue, Q. Xie, Q. Zhao, Y . Zheng, and D. Meng, “From rain generation to rain removal,” inCVPR, 2021

    work page 2021

  61. [61]

    Un- supervised deraining: Where contrastive learning meets self-similarity,

    Y . Ye, C. Yu, Y . Chang, L. Zhu, X.-l. Zhao, L. Yan, and Y . Tian, “Un- supervised deraining: Where contrastive learning meets self-similarity,” inCVPR, 2022

    work page 2022

  62. [62]

    Memory oriented transfer learning for semi-supervised image deraining,

    H. Huang, A. Yu, and R. He, “Memory oriented transfer learning for semi-supervised image deraining,” inCVPR, 2021

    work page 2021

  63. [63]

    Towards robust rain removal against adversarial attacks: A comprehensive benchmark analysis and beyond,

    Y . Yu, W. Yang, Y .-P. Tan, and A. C. Kot, “Towards robust rain removal against adversarial attacks: A comprehensive benchmark analysis and beyond,” inCVPR, 2022

    work page 2022

  64. [64]

    Deep learning for seeing through window with raindrops,

    Y . Quan, S. Deng, Y . Chen, and H. Ji, “Deep learning for seeing through window with raindrops,” inICCV, 2019

    work page 2019

  65. [65]

    Removing raindrops and rain streaks in one go,

    R. Quan, X. Yu, Y . Liang, and Y . Yang, “Removing raindrops and rain streaks in one go,” inCVPR, 2021

    work page 2021

  66. [66]

    Rain or snow detection in image sequences through use of a histogram of orientation of streaks,

    J. Bossu, N. Hautiere, and J.-P. Tarel, “Rain or snow detection in image sequences through use of a histogram of orientation of streaks,”IJCV, 2011

    work page 2011

  67. [67]

    Frequency-oriented efficient transformer for all-in-one weather-degraded image restoration,

    T. Gao, Y . Wen, K. Zhang, J. Zhang, T. Chen, L. Liu, and W. Luo, “Frequency-oriented efficient transformer for all-in-one weather-degraded image restoration,”Transactions on Circuits and Sys- tems for Video Technology, 2023

    work page 2023

  68. [68]

    Gridformer: Residual dense transformer with grid structure for image restoration in adverse weather conditions,

    T. Wang, K. Zhang, Z. Shao, W. Luo, B. Stenger, T. Lu, T.-K. Kim, W. Liu, and H. Li, “Gridformer: Residual dense transformer with grid structure for image restoration in adverse weather conditions,”IJCV, 2024

    work page 2024

  69. [69]

    Mwformer: Multi- weather image restoration using degradation-aware transformers,

    R. Zhu, Z. Tu, J. Liu, A. C. Bovik, and Y . Fan, “Mwformer: Multi- weather image restoration using degradation-aware transformers,”TIP, 2024

    work page 2024

  70. [70]

    Towards real-world adverse weather image restoration: Enhancing clearness and semantics with vision-language models,

    J. Xu, M. Wu, X. Hu, C.-W. Fu, Q. Dou, and P.-A. Heng, “Towards real-world adverse weather image restoration: Enhancing clearness and semantics with vision-language models,” inECCV, 2024

    work page 2024

  71. [71]

    Blind image decomposition,

    J. Han, W. Li, P. Fang, C. Sun, J. Hong, M. A. Armin, L. Petersson, and H. Li, “Blind image decomposition,” inECCV, 2022

    work page 2022

  72. [72]

    Rain streak removal using layer priors,

    Y . Li, R. T. Tan, X. Guo, J. Lu, and M. S. Brown, “Rain streak removal using layer priors,” inCVPR, 2016

    work page 2016

  73. [73]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” inNeurIPS, 2017

    work page 2017

  74. [74]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” inMICCAI, 2015

    work page 2015

  75. [75]

    High- resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High- resolution image synthesis with latent diffusion models,” inCVPR, 2022

    work page 2022

  76. [76]

    Twins: Revisiting the design of spatial attention in vision transformers,

    X. Chu, Z. Tian, Y . Wang, B. Zhang, H. Ren, X. Wei, H. Xia, and C. Shen, “Twins: Revisiting the design of spatial attention in vision transformers,” inNeurIPS, 2021

    work page 2021

  77. [77]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” inICCV, 2021

    work page 2021

  78. [78]

    Swin transformer v2: Scaling up capacity and resolution,

    Z. Liu, H. Hu, Y . Lin, Z. Yao, Z. Xie, Y . Wei, J. Ning, Y . Cao, Z. Zhang, L. Donget al., “Swin transformer v2: Scaling up capacity and resolution,” inCVPR, 2022

    work page 2022

  79. [79]

    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,” inCVPR, 2022

    work page 2022

  80. [80]

    Simple baselines for image restoration,

    L. Chen, X. Chu, X. Zhang, and J. Sun, “Simple baselines for image restoration,” inECCV, 2022

    work page 2022

Showing first 80 references.