ShadowHack: Hacking Shadows via Luminance-Color Divide and Conquer

Jin Hu; Mingjia Li; Xiaojie Guo

REVIEW 3 major objections 7 minor 1 cited by

ShadowHack: Hacking Shadows via Luminance-Color Divide and Conquer

T0 review · 3 major / 7 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read ShadowHack claims that shadow removal is best done by first restoring luminance and then regenerating color, and it reports state-of-the-art PSNR and RMSE on the ISTD+ and SRD benchmarks.

desk verdict A solid two-stage shadow removal with strong ablations, but the headline SOTA claim rests on an unspecified mask pathway that needs clarification or code. read the letter →

arxiv 2412.02545 v3 pith:X25GHJOY submitted 2024-12-03 cs.CV

Jin Hu , Mingjia Li , Xiaojie Guo This is my paper · ORCID

classification cs.CV

keywords shadowremovalluminancerestorationcolorregenerationrectifiedoutreachattentioncross-attentionYCbCrdecompositionimagetransformer

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 tries to establish that shadow removal can be solved more effectively by decomposing the problem into two sequential subproblems: luminance recovery and color regeneration. The authors argue that shadows degrade images in three entangled ways—reduced brightness, texture deterioration, and color distortion—and that a divide-and-conquer pipeline handles each more cleanly than a single end-to-end model. On standard benchmarks, the proposed two-stage network, ShadowHack, reports the best all-image PSNR and RMSE, including an RMSE improvement of 0.43 on SRD over the prior best. A sympathetic reader would care because the result suggests that explicit decomposition into luminance and chrominance is a viable design principle for shadow-removal networks, not just a pre-processing trick.

What carries the argument

The load-bearing machinery is the YCbCr decoupling plus two specialized networks. LRNet is a U-shaped transformer whose deeper stages replace local blocks with a Rectified Outreach Attention (ROA) module: queries come from regular window partitions, keys and values come from dilated outreach windows, and the final attention is a rectified difference between an attention map computed on luminance-plus-color features and one computed on color-only features, which suppresses noise and negative correlation in shadow areas. CRNet is a U-shaped network augmented with a pretrained color encoder; cross-attention modules use restored luminance features as queries and keys and multi-scale color features as values to inject colors into the skip connections. A checkpoint-ensembling trick used only during training makes CRNet robust to imperfect luminance outputs without adding inference cost.

What would settle it

Construct or find an image in which a shadowed object and a lit object have the same measured luminance but very different true color, such as a dark-red object in shadow and a bright-blue object in light. If ShadowHack turns the shadowed object blue, the luminance-as-material assumption is violated; if it keeps the object red, the network is using information beyond luminance similarity. A quantitative version would measure CRNet's output color error on synthetic shadow patches whose matched lit counterparts have identical luminance but different chroma.

Watch

Extended reading notes

Core claim

The central discovery is that separating shadow removal into luminance restoration and color regeneration—rather than estimating a single shadow-free RGB image—lets each sub-network work from cleaner inputs and produces more accurate results. The pipeline converts RGB to YCbCr, restores the Y channel with LRNet, then regenerates the Cb and Cr channels in CRNet by using restored luminance features as queries and keys and color features as values, so colors in shadow regions are copied from non-shadow patches with similar luminance. The paper reports an all-image PSNR of 36.31 and RMSE of 2.48 on ISTD+, and 35.94 and 2.90 on SRD, surpassing the previous best results on most metrics. Ablation experiments show that the decomposition itself, the outreach windows, the differential rectification, and the cross-attention color injection each contribute to the gains.

Load-bearing premise

The color regeneration network assumes that two patches with similar luminance are made of similar material, so it copies color from non-shadow patches whose brightness matches the shadow region; if that assumption fails, the wrong color is injected.

Editorial extensions

If this is right

If the reported numbers hold, end-to-end RGB shadow-removal models are not the only strong option; decomposing into luminance and color is a competitive design that can be combined with future backbones.
The SRD RMSE gain of 0.43 suggests color regeneration is where decoupling pays off most, so methods that ignore color bias will keep lagging on color-error metrics.
The Rectified Outreach Attention module could be dropped into other window-based restoration transformers whenever a region mask is available.
Because CRNet copies colors from non-shadow patches by luminance similarity, its accuracy should degrade gracefully as shadow masks become coarser, provided the mask-refine network keeps working.
The two-stage pipeline makes shadow removal user-controllable: users can supply or edit the shadow mask, and only the luminance correction and the corresponding color-copy region change.

Reading between the lines

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

The paper does not test what happens when luminance similarity fails to indicate material identity; an image with two objects of equal brightness but different true colors in shadow and lit regions would probe this boundary.
The same YCbCr divide-and-conquer design may transfer to low-light enhancement or shadow detection, where the luminance channel carries most of the illumination problem; applying LRNet alone to underexposed images would be a direct test.
The checkpoint-ensembling trick suggests CRNet is sensitive to training-time luminance errors; feeding deliberately corrupted luminance maps at test time would quantify that sensitivity.
If the public code reproduces the reported numbers, the strongest next comparison is against diffusion-based shadow removal at matched parameter counts, since ShadowHack claims a large RMSE improvement with a smaller, non-diffusion model.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a solid shadow-removal paper with a real architectural idea and unusually complete ablations, but the headline SOTA numbers are less certain than the text implies because the inference-time mask pathway is never specified. If you are going to rely on these numbers, ask for code or a precise description of the pipeline.

What's new: the luminance-color decoupling in YCbCr, with LRNet handling brightness/texture and CRNet doing color regeneration via cross-attention from luminance features to color features. The rectified outreach attention is a reasonable tweak on window attention, and the differential rectification is well motivated. The ablations in Tables 3-5 are genuinely informative; they isolate the contributions of decoupling, outreach, rectification, and cross-attention cleanly.

The SRD results are the more convincing part of the empirical claim: RMSE 2.90 vs 3.33 for the next best, a solid margin. On ISTD+, the all-image PSNR gain is 0.15 dB over RASM, which is within normal run-to-run variation for this kind of model. The absence of error bars or repeated runs is a real limitation.

Now the soft spot, and it is a real one. The main architecture in Sec. 4 feeds only luminance and color into LRNet and CRNet; no shadow mask is shown as an input. Yet Sec. 5.4 shows that user-specified masks change the output and introduces a mask-refine U-Net. The paper never states whether that mask-refine U-Net is applied during the Table 1 evaluations. Since SRD masks are predicted by DHAN and could be noisy, a hidden mask-refinement stage would give ShadowHack cleaner mask inputs than the baselines, and part of the reported gain could come from mask preprocessing rather than the proposed decoupling. This is a verifiability problem, not necessarily a flaw: the authors may be using the raw masks in Table 1, but they need to say so clearly or release code.

The luminance-as-material-proxy assumption in CRNet is a legitimate theoretical concern, but the checkpoint ensemble shows the authors are aware of the fragility, and the ablations support the design. Minor point: the physics analysis in Sec. 3 is heuristic and not load-bearing.

This paper deserves a serious referee. The method is novel enough, the ablations are strong, and the main ambiguity is addressable with a clear statement or code. A reviewer should focus on the exact inference pipeline for the benchmark numbers and the robustness of the SOTA claim.

Referee Report

3 major / 7 minor

Summary. ShadowHack proposes a two-stage divide-and-conquer framework for single-image shadow removal. An invertible decoupling maps each RGB input into a luminance channel It and color channels Ic; LRNet restores brightness and texture in the luminance space using a Rectified Outreach Attention module with dilated overlapping windows and a differential attention-rectification mechanism, while CRNet regenerates color through cross-attention in which restored luminance features serve as query and key and multi-scale color features from a ConvNext-v2 color encoder serve as value, with training-time checkpoint ensembling to tolerate imperfect LRNet outputs. Experiments on ISTD+ and SRD report state-of-the-art numbers (PSNR 36.31 and RMSE 2.48 on ISTD+; PSNR 35.94 and RMSE 2.90 on SRD), along with ablations of the decoupling, the attention module, and the color network, and qualitative generalization results on UIUC and UCF. The paper states that code will be released.

Significance. The luminance-color decomposition is well motivated by the image-formation analysis in Sec. 3 (Eqs. 1-7), and the ablation study is a genuine strength: each design choice, including decoupling versus end-to-end RGB processing, outreach dilation, rectification, feature combinations in the attention module, cross-attention versus concatenation in CRNet, and checkpoint ensembling, is backed by a quantitative table (Tables 3-5) with internally consistent conclusions. The SRD result (ALL RMSE 2.90 versus 3.33 for the prior best) is a comparatively large improvement, the model is lightweight (23.3M parameters), and the re-evaluation of RASM from released output images is a fair comparison practice. The significance of the headline SOTA claim is, however, currently limited: the inference-time role of the shadow mask is never specified, so the SRD gains cannot yet be attributed to the proposed architecture rather than to unseen mask preprocessing, and the ISTD+ margin (0.15 dB) is reported without any error analysis. If the mask pathway is clarified and the gains hold up, this will be a solid and clean contribution.

major comments (3)

[Sec. 4 / Fig. 4; Sec. 5.4] The methodology section never specifies how the shadow mask enters the network. Eqs. (8)-(9) and Fig. 4 define the pipeline as D^{-1}(C(R(It), Ic)) with no mask term, yet Sec. 5.4 (Fig. 11) demonstrates that user-specified masks change the output at inference, and the same section introduces a mask-refine U-Net to 'get rid of the performance degradation brought by inaccurate masks.' The paper must state (a) the exact point at which the mask conditions LRNet or CRNet, (b) whether the mask-refine network was applied in the Table 1 evaluations, especially for SRD where masks come from DHAN predictions, and (c) the mask-refine network's training data, supervision, and loss. Without this information, the reported SRD improvements (ALL RMSE 2.90 vs 3.33; shadow-region PSNR 39.47 vs 38.81 over Homoformer) could stem from mask refinement rather than from the luminance-color divide-and-conquer, and the headline SOTA claim is not verifiable from the manuscript.
[Sec. 5.2 / Table 1] The ISTD+ state-of-the-art claim rests on a 0.15 dB PSNR margin (36.31 vs 36.16 for RASM) and a 0.05 RMSE margin (2.48 vs 2.53), reported without error bars, repeated-run statistics, or any significance analysis. The claim also depends on region aggregation: ShadowHack's shadow-region metrics on ISTD+ (PSNR 40.56, RMSE 4.46) are slightly worse than RASM's (40.73, 4.41), so the advantage is concentrated in non-shadow regions. Given the unspecified mask pathway raised above, I ask for per-image error distributions (or equivalent statistical support) and a discussion of the shadow-region deficit before the paper can fairly claim to surpass the prior art.
[Sec. 4.2 / Table 5] The color regeneration mechanism rests on the assumption that luminance similarity is a reliable proxy for material identity: CRNet's cross-attention computes query and key from restored luminance features while color features serve as values, and the text states the aim is to 'reference regions with similar/identical material properties.' If a shadow region and a lit region have similar brightness but different true albedo, the attention will copy incorrect colors. The paper neither analyzes this failure mode nor tests the assumption beyond the aggregate Table 5 ablation; the checkpoint-ensemble paragraph in the same section implicitly concedes that LRNet's test-time outputs are imperfect, which heightens the concern. I ask for an analysis of this failure mode (for example, attention visualizations or a targeted study on same-luminance/different-albedo pairs), or an explicit statement of the limitation.

minor comments (7)

[Table 3 vs Table 1] The ablation in Table 3 reports an ALL RMSE of 2.46 for 'Ours' on ISTD+, while Table 1 reports 2.48 for ShadowHack under the same protocol; please reconcile these numbers.
[Table 3] The RGB-input baseline in Table 3 attains PSNR 36.16 on ISTD+, exactly RASM's ALL PSNR in Table 1; please clarify whether this baseline is an independently trained model or a re-implementation or reuse of RASM.
[Sec. 5.1] AdamW has no 'momentum' hyperparameter; the values (0.9, 0.999) are the beta_1 and beta_2 parameters of the optimizer, and the text should be corrected accordingly.
[Sec. 4.2] The checkpoint-ensemble description is ambiguous: the sentence 'the luminance network often yields better performance on the training set' does not state the reference point of the comparison, and the number and selection scheme of the early-stage checkpoints are not given; please specify how many checkpoints are used and how they are sampled.
[Abstract] The abstract states that 'Our code will be made publicly available at here,' with a placeholder instead of a URL; please complete the link or remove the placeholder.
[Sec. 3] The phrase 'To kindle a poorly lit image' appears to be a word-choice error; 'kindle' should likely be 'brighten' or 'enhance.'
[Eq. (13)] The notation lambda_j^i with i, j in {1, 2} is hard to parse; please give the four scalar parameters distinct, readable names.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the SOTA claim is evaluated on external ISTD+/SRD test sets, and the only self-citation (RASM) is a re-evaluation of released output images, not a load-bearing derivation.

full rationale

The central claim is empirical benchmark performance on external test sets, reported in Sec. 5.2. ShadowHack's pipeline (Eqs. 8-12, Sec. 4.1-4.2) is a learned luminance/color decomposition trained with L1 and VGG losses; no parameter is fitted to the test quantity and then renamed a prediction. The comparison against the authors' own RASM is transparently a re-evaluation of its official released outputs, which is independent evidence rather than a self-citation chain. The robustness section's mask-refine network (Sec. 5.4, Fig. 12) is under-specified for the main tables, and whether it is applied in the Table 1 evaluations is a reproducibility gap, but no equation reduces the final PSNR/RMSE to a fitted input or to the mask-refine design by construction. The luminance-as-material-proxy assumption in CRNet is a correctness risk, not a circularity. Overall no significant circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The framework is an engineering combination of standard image formation, YCbCr decomposition, transformer attention, and pretrained encoders; no new physical entities are introduced. The main hand-chosen constants are lambda_0 and the outreach overlap ratio.

free parameters (2)

lambda_0 in rectified attention = 0.7
Predefined constant in Eq. (13) controlling the rectification offset, chosen by hand and reported in Sec. 5.1.
Outreach window overlapping ratio = 0.5
Ratio for dilated outreach window partitions in ROA, set in Sec. 5.1.

assumptions (4)

domain assumption Radiometric image formation model with a single dominant light source and shadow as attenuated ambient term
Used to derive that shadow regions suffer reduced brightness and color bias; assumes no inter-reflections or colored ambient light (Sec. 3, Eqs. 1-7).
ad hoc to paper YCbCr luminance/chrominance separation isolates brightness and texture from color bias
Motivated by color bias plots in Fig. 3, but it is a heuristic design choice rather than a proven decomposition (Sec. 4).
domain assumption Pretrained ConvNext-v2 on ImageNet-21k transfers useful color features for shadow colorization
CRNet's color encoder relies on ImageNet-21k pretraining and assumes those features generalize to shadow colorization (Sec. 4.2).
domain assumption Non-shadow regions provide sufficient color references for shadow regions
Exemplar-based colorization premise; if a scene has unique materials only in shadow, color recovery is underdetermined (Sec. 4.2).

how reviews work

0 comments

Cite this review

Pith. "Pith review of ShadowHack: Hacking Shadows via Luminance-Color Divide and Conquer." pith.science (2026). https://pith.science/paper/X25GHJOY

@misc{pith2026241202545,
  author       = {Pith},
  title        = {Pith review of: ShadowHack: Hacking Shadows via Luminance-Color Divide and Conquer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/X25GHJOY}},
  note         = {Machine review of arXiv:2412.02545}
}

read the original abstract

Shadows introduce challenges such as reduced brightness, texture deterioration, and color distortion in images, complicating a holistic solution. This study presents \textbf{ShadowHack}, a divide-and-conquer strategy that tackles these complexities by decomposing the original task into luminance recovery and color remedy. To brighten shadow regions and repair the corrupted textures in the luminance space, we customize LRNet, a U-shaped network with a rectified attention module, to enhance information interaction and recalibrate contaminated attention maps. With luminance recovered, CRNet then leverages cross-attention mechanisms to revive vibrant colors, producing visually compelling results. Extensive experiments on multiple datasets are conducted to demonstrate the superiority of ShadowHack over existing state-of-the-art solutions both quantitatively and qualitatively, highlighting the effectiveness of our design. Our code will be made publicly available.

Figures

Figures reproduced from arXiv: 2412.02545 by the authors.

**Figure 2.** Example of the complex degradation in a shadow image. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Color bias in shadow regions between the reflectances of [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (7 more)

**Figure 4.** Figure 4: The overall architecture of ShadowHack comprises two core components: (a) a luminance restoration network, which focuses on [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

**Figure 6.** Figure 6: An example of our rectified outreach attention. The query [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]

**Figure 7.** Figure 7: Visual comparison on a sample from the ISTD+ dataset. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

**Figure 8.** Figure 8: Visual comparison on a sample from the SRD dataset. [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

**Figure 9.** Figure 9: Ablation study on rectified outreach attention. The results [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]

**Figure 10.** Figure 10: Visual comparisons on the UIUC and UCF datasets. [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]

**Figure 11.** Figure 11: Visual results of selective shadow masks on the UCF dataset. [PITH_FULL_IMAGE:figures/full_fig_p009_11.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

NTIRE 2025 Image Shadow Removal Challenge Report
cs.CV 2025-06 conditional novelty 4.0 of 10

The NTIRE 2025 shadow removal challenge report gives a leaderboard of 17 methods on the WSRD+ dataset and a data alignment upgrade that raises baseline PSNR by about 2 dB.

Reference graph

Works this paper leans on

58 extracted references · 53 canonical work pages · cited by 1 Pith paper

[1]

Recovering intrinsic scene characteristics

Harry Barrow, J Tenenbaum, A Hanson, and E Riseman. Recovering intrinsic scene characteristics. Comput. vis. syst, 2(3-26):2, 1978. 1, 3

work page 1978
[2]

Retinexformer: One-stage retinex- based transformer for low-light image enhancement

Yuanhao Cai, Hao Bian, Jing Lin, Haoqian Wang, Radu Tim- ofte, and Yulun Zhang. Retinexformer: One-stage retinex- based transformer for low-light image enhancement. InICCV, pages 12470–12479, 2023. 1

work page 2023
[3]

Emerg- ing properties in self-supervised vision transformers

Mathilde Caron, Hugo Touvron, Ishan Misra, Herv´e J´egou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerg- ing properties in self-supervised vision transformers. In ICCV, pages 9630–9640, 2021. 3

work page 2021
[4]

Pre-trained image processing transformer

Hanting Chen, Yunhe Wang, Tianyu Guo, Chang Xu, Yiping Deng, Zhenhua Liu, Siwei Ma, Chunjing Xu, Chao Xu, and Wen Gao. Pre-trained image processing transformer. In CVPR, pages 12299–12310, 2021. 3

work page 2021
[5]

A comparative study of image restoration networks for general backbone network design

Xiangyu Chen, Zheyuan Li, Yuandong Pu, Yihao Liu, Jiantao Zhou, Yu Qiao, and Chao Dong. A comparative study of image restoration networks for general backbone network design. arXiv preprint arXiv:2310.11881, 2023. 3, 5

arXiv 2023
[6]

Activating more pixels in image super-resolution transformer

Xiangyu Chen, Xintao Wang, Jiantao Zhou, Yu Qiao, and Chao Dong. Activating more pixels in image super-resolution transformer. In CVPR, pages 22367–22377, 2023. 3, 5

work page 2023
[7]

Blue shadows: physical, physiological, and psychological causes

Michael E Churma. Blue shadows: physical, physiological, and psychological causes. Applied optics, 33(21):4719–4722,

work page
[8]

Towards ghost- free shadow removal via dual hierarchical aggregation net- work and shadow matting GAN

Xiaodong Cun, Chi-Man Pun, and Cheng Shi. Towards ghost- free shadow removal via dual hierarchical aggregation net- work and shadow matting GAN. In AAAI, pages 10680– 10687, 2020. 6, 7, 8

work page 2020

Show all 58 references

[9]

An image is worth 16x16 words: Transformers for image recognition at scale

Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

2021
[10]

Finlayson, Steven D

Graham D. Finlayson, Steven D. Hordley, and Mark S. Drew. Removing shadows from images. In ECCV (4), pages 823– 836, 2002. 2

2002
[11]

Finlayson, Steven D

Graham D. Finlayson, Steven D. Hordley, Cheng Lu, and Mark S. Drew. On the removal of shadows from images. IEEE Trans. Pattern Anal. Mach. Intell., 28(1):59–68, 2006. 2

2006
[12]

Finlayson, Mark S

Graham D. Finlayson, Mark S. Drew, and Cheng Lu. Entropy minimization for shadow removal. Int. J. Comput. Vis., 85 (1):35–57, 2009. 2

2009
[13]

Shadowformer: Global context helps shadow removal

Lanqing Guo, Siyu Huang, Ding Liu, Hao Cheng, and Bihan Wen. Shadowformer: Global context helps shadow removal. In AAAI, pages 710–718, 2023. 1, 3, 6, 7, 8

2023
[14]

Shadowdiffusion: When degradation prior meets diffusion model for shadow removal

Lanqing Guo, Chong Wang, Wenhan Yang, Siyu Huang, Yufei Wang, Hanspeter Pfister, and Bihan Wen. Shadowdiffusion: When degradation prior meets diffusion model for shadow removal. In CVPR, pages 14049–14058, 2023. 1, 2, 6, 7, 8

2023
[15]

Paired regions for shadow detection and removal

Ruiqi Guo, Qieyun Dai, and Derek Hoiem. Paired regions for shadow detection and removal. IEEE Trans. Pattern Anal. Mach. Intell., 35(12):2956–2967, 2013. 1, 2, 7, 9

2013
[16]

Low-light image enhancement via breaking down the darkness

Xiaojie Guo and Qiming Hu. Low-light image enhancement via breaking down the darkness. Int. J. Comput. Vis., 131(1): 48–66, 2023. 1, 8

2023
[17]

LIME: low-light image enhancement via illumination map estimation

Xiaojie Guo, Yu Li, and Haibin Ling. LIME: low-light image enhancement via illumination map estimation. IEEE Trans. Image Process., 26(2):982–993, 2017. 1, 2, 3

2017
[18]

Neighborhood attention transformer

Ali Hassani, Steven Walton, Jiachen Li, Shen Li, and Humphrey Shi. Neighborhood attention transformer. In CVPR, pages 6185–6194, 2023. 3

2023
[19]

Shadow elimination for effective moving object detection by gaussian shadow modeling

Jun-Wei Hsieh, Wen-Fong Hu, Chia-Jung Chang, and Yung- Sheng Chen. Shadow elimination for effective moving object detection by gaussian shadow modeling. Image Vis. Comput., 21(6):505–516, 2003. 1

2003
[20]

Mask-shadowgan: Learning to remove shadows from unpaired data

Xiaowei Hu, Yitong Jiang, Chi-Wing Fu, and Pheng-Ann Heng. Mask-shadowgan: Learning to remove shadows from unpaired data. In ICCV, pages 2472–2481, 2019. 1, 2

2019
[21]

Unicolor: A unified framework for multi-modal colorization with trans- former

Zhitong Huang, Nanxuan Zhao, and Jing Liao. Unicolor: A unified framework for multi-modal colorization with trans- former. ACM Trans. Graph., 41(6):205:1–205:16, 2022. 6

2022
[22]

Yeying Jin, Wei Ye, Wenhan Yang, Yuan Yuan, and Robby T. Tan. Des3: Adaptive attention-driven self and soft shadow removal using vit similarity. In AAAI, pages 2634–2642,
[23]

Automatic shadow detection and removal from a single image

Salman Hameed Khan, Mohammed Bennamoun, Fer- dous Ahmed Sohel, and Roberto Togneri. Automatic shadow detection and removal from a single image. IEEE Trans. Pattern Anal. Mach. Intell., 38(3):431–446, 2016. 2

2016
[24]

The retinex theory of color vision

Edwin H Land. The retinex theory of color vision. Scientific american, 237(6):108–129, 1977. 1, 2

1977
[25]

Shadow removal via shadow image decomposition

Hieu Le and Dimitris Samaras. Shadow removal via shadow image decomposition. In ICCV, pages 8577–8586, 2019. 1, 2, 7

2019
[26]

Physics-based shadow image decomposition for shadow removal

Hieu Le and Dimitris Samaras. Physics-based shadow image decomposition for shadow removal. IEEE Trans. Pattern Anal. Mach. Intell., 44(12):9088–9101, 2022. 1, 2, 3

2022
[27]

Reference-based sketch image colorization using augmented-self reference and dense semantic correspondence

Junsoo Lee, Eungyeup Kim, Yunsung Lee, Dongjun Kim, Jaehyuk Chang, and Jaegul Choo. Reference-based sketch image colorization using augmented-self reference and dense semantic correspondence. In CVPR, pages 5800–5809, 2020. 6

2020
[28]

Tsang, and Song Wang

Xiaoguang Li, Qing Guo, Rabab Abdelfattah, Di Lin, Wei Feng, Ivor W. Tsang, and Song Wang. Leveraging inpainting for single-image shadow removal. In ICCV, pages 13009– 13018, 2023. 1, 2, 6, 7, 8

2023
[29]

Regional atten- tion for shadow removal

Hengxing Liu, Mingjia Li, and Xiaojie Guo. Regional atten- tion for shadow removal. In ACM Multimedia, 2024. 1, 3, 6, 7, 8

2024
[30]

A decoupled multi-task network for shadow removal

Jiawei Liu, Qiang Wang, Huijie Fan, Wentao Li, Liangqiong Qu, and Yandong Tang. A decoupled multi-task network for shadow removal. IEEE Trans. Multim., 25:9449–9463, 2023. 7

2023
[31]

Yuhao Liu, Zhanghan Ke, Ke Xu, Fang Liu, Zhenwei Wang, and Rynson W. H. Lau. Recasting regional lighting for shadow removal. In AAAI, pages 3810–3818, 2024. 1, 2, 6, 8 10

2024
[32]

Swin transformer: Hierarchical vision transformer using shifted windows

Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In ICCV, pages 9992–10002, 2021. 3

2021
[33]

From shadow generation to shadow removal

Zhihao Liu, Hui Yin, Xinyi Wu, Zhenyao Wu, Yang Mi, and Song Wang. From shadow generation to shadow removal. In CVPR, pages 4927–4936, 2021. 1, 2, 6, 8

2021
[34]

A convnet for the 2020s

Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feicht- enhofer, Trevor Darrell, and Saining Xie. A convnet for the 2020s. In CVPR, pages 11966–11976, 2022. 6, 9

2022
[35]

Decoupled weight decay regularization

Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. In ICLR, 2019. 6

2019
[36]

Gray2colornet: Transfer more colors from reference image

Peng Lu, Jinbei Yu, Xujun Peng, Zhaoran Zhao, and Xiaojie Wang. Gray2colornet: Transfer more colors from reference image. In ACM Multimedia, pages 3210–3218, 2020. 6

2020
[37]

David K. Lynch. Shadows. Appl. Opt., 54(4):B154–B164,
[38]

Physical models for moving shadow and object detection in video

Sohail Nadimi and Bir Bhanu. Physical models for moving shadow and object detection in video. IEEE Trans. Pattern Anal. Mach. Intell., 26(8):1079–1087, 2004. 1

2004
[39]

Liangqiong Qu, Jiandong Tian, Shengfeng He, Yandong Tang, and Rynson W. H. Lau. Deshadownet: A multi-context em- bedding deep network for shadow removal. In CVPR, pages 2308–2316, 2017. 1, 2, 7

2017
[40]

U-net: Convolutional networks for biomedical image segmentation

Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In MICCAI (3), pages 234–241. Springer, 2015. 8, 9

2015
[41]

The shadow meets the mask: Pyramid-based shadow removal

Yael Shor and Dani Lischinski. The shadow meets the mask: Pyramid-based shadow removal. Comput. Graph. Forum, 27 (2):577–586, 2008. 2, 3

2008
[42]

Weiss, Niru Maheswaranathan, and Surya Ganguli

Jascha Sohl-Dickstein, Eric A. Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics. In ICML, pages 2256– 2265, 2015. 2

2015
[43]

Style-guided shadow removal

Jin Wan, Hui Yin, Zhenyao Wu, Xinyi Wu, Yanting Liu, and Song Wang. Style-guided shadow removal. In ECCV (19), pages 361–378, 2022. 1, 6, 7, 8

2022
[44]

Stacked conditional generative adversarial networks for jointly learning shadow detection and shadow removal

Jifeng Wang, Xiang Li, and Jian Yang. Stacked conditional generative adversarial networks for jointly learning shadow detection and shadow removal. In CVPR, pages 1788–1797,
[45]

Uformer: A general u-shaped transformer for image restoration

Zhendong Wang, Xiaodong Cun, Jianmin Bao, Wengang Zhou, Jianzhuang Liu, and Houqiang Li. Uformer: A general u-shaped transformer for image restoration. In CVPR, pages 17662–17672, 2022. 3, 5

2022
[46]

A bayesian approach for shadow extraction from a single image

Tai-Pang Wu and Chi-Keung Tang. A bayesian approach for shadow extraction from a single image. In ICCV, pages 480–487, 2005. 2

2005
[47]

Vision transformer with deformable attention

Zhuofan Xia, Xuran Pan, Shiji Song, Li Erran Li, and Gao Huang. Vision transformer with deformable attention. In CVPR, pages 4784–4793, 2022. 3

2022
[48]

Homoformer: Homogenized trans- former for image shadow removal

Jie Xiao, Xueyang Fu, Yurui Zhu, Dong Li, Jie Huang, Kai Zhu, and Zheng-Jun Zha. Homoformer: Homogenized trans- former for image shadow removal. In CVPR, pages 25617– 25626, 2024. 1, 3, 6, 7, 8

2024
[49]

Cretinex: A progressive color-shift aware retinex model for low-light image enhancement

Han Xu, Hao Zhang, Xunpeng Yi, and Jiayi Ma. Cretinex: A progressive color-shift aware retinex model for low-light image enhancement. Int. J. Comput. Vis., 132(9):3610–3632,
[50]

Shadow removal using bilateral filtering

Qingxiong Yang, Kar-Han Tan, and Narendra Ahuja. Shadow removal using bilateral filtering. IEEE Trans. Image Process., 21(10):4361–4368, 2012. 2

2012
[51]

Differential transformer.arXiv preprint arXiv:2410.05258, 2024

Tianzhu Ye, Li Dong, Yuqing Xia, Yutao Sun, Yi Zhu, Gao Huang, and Furu Wei. Differential transformer.arXiv preprint arXiv:2410.05258, 2024. 5

2024 arXiv
[52]

Restormer: Efficient transformer for high-resolution image restoration

Syed Waqas Zamir, Aditya Arora, Salman Khan, Mu- nawar Hayat, Fahad Shahbaz Khan, and Ming-Hsuan Yang. Restormer: Efficient transformer for high-resolution image restoration. In CVPR, pages 5718–5729, 2022. 3, 5

2022
[53]

Shadow re- mover: Image shadow removal based on illumination recover- ing optimization

Ling Zhang, Qing Zhang, and Chunxia Xiao. Shadow re- mover: Image shadow removal based on illumination recover- ing optimization. IEEE Trans. Image Process., 24(11):4623– 4636, 2015. 2

2015
[54]

RIS-GAN: explore residual and illumination with gen- erative adversarial networks for shadow removal

Ling Zhang, Chengjiang Long, Xiaolong Zhang, and Chunxia Xiao. RIS-GAN: explore residual and illumination with gen- erative adversarial networks for shadow removal. In AAAI, pages 12829–12836, 2020. 1, 2

2020
[55]

Improving shadow suppression for illumination robust face recognition

Wuming Zhang, Xi Zhao, Jean-Marie Morvan, and Liming Chen. Improving shadow suppression for illumination robust face recognition. IEEE Trans. Pattern Anal. Mach. Intell., 41 (3):611–624, 2019. 1

2019
[56]

Wei Zhou, Julie Stephany Berrio, Stewart Worrall, and Ed- uardo M. Nebot. Automated evaluation of semantic segmen- tation robustness for autonomous driving. IEEE Trans. Intell. Transp. Syst., 21(5):1951–1963, 2020. 1

1951
[57]

Jiejie Zhu, Kegan G. G. Samuel, Syed Zain Masood, and Mar- shall F. Tappen. Learning to recognize shadows in monochro- matic natural images. In CVPR, pages 223–230, 2010. 7, 9

2010
[58]

Bijective mapping network for shadow removal

Yurui Zhu, Jie Huang, Xueyang Fu, Feng Zhao, Qibin Sun, and Zheng-Jun Zha. Bijective mapping network for shadow removal. In CVPR, pages 5617–5626, 2022. 2, 6, 7, 8 11

2022

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Recovering intrinsic scene characteristics

Harry Barrow, J Tenenbaum, A Hanson, and E Riseman. Recovering intrinsic scene characteristics. Comput. vis. syst, 2(3-26):2, 1978. 1, 3

work page 1978

[2] [2]

Retinexformer: One-stage retinex- based transformer for low-light image enhancement

Yuanhao Cai, Hao Bian, Jing Lin, Haoqian Wang, Radu Tim- ofte, and Yulun Zhang. Retinexformer: One-stage retinex- based transformer for low-light image enhancement. InICCV, pages 12470–12479, 2023. 1

work page 2023

[3] [3]

Emerg- ing properties in self-supervised vision transformers

Mathilde Caron, Hugo Touvron, Ishan Misra, Herv´e J´egou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerg- ing properties in self-supervised vision transformers. In ICCV, pages 9630–9640, 2021. 3

work page 2021

[4] [4]

Pre-trained image processing transformer

Hanting Chen, Yunhe Wang, Tianyu Guo, Chang Xu, Yiping Deng, Zhenhua Liu, Siwei Ma, Chunjing Xu, Chao Xu, and Wen Gao. Pre-trained image processing transformer. In CVPR, pages 12299–12310, 2021. 3

work page 2021

[5] [5]

A comparative study of image restoration networks for general backbone network design

Xiangyu Chen, Zheyuan Li, Yuandong Pu, Yihao Liu, Jiantao Zhou, Yu Qiao, and Chao Dong. A comparative study of image restoration networks for general backbone network design. arXiv preprint arXiv:2310.11881, 2023. 3, 5

arXiv 2023

[6] [6]

Activating more pixels in image super-resolution transformer

Xiangyu Chen, Xintao Wang, Jiantao Zhou, Yu Qiao, and Chao Dong. Activating more pixels in image super-resolution transformer. In CVPR, pages 22367–22377, 2023. 3, 5

work page 2023

[7] [7]

Blue shadows: physical, physiological, and psychological causes

Michael E Churma. Blue shadows: physical, physiological, and psychological causes. Applied optics, 33(21):4719–4722,

work page

[8] [8]

Towards ghost- free shadow removal via dual hierarchical aggregation net- work and shadow matting GAN

Xiaodong Cun, Chi-Man Pun, and Cheng Shi. Towards ghost- free shadow removal via dual hierarchical aggregation net- work and shadow matting GAN. In AAAI, pages 10680– 10687, 2020. 6, 7, 8

work page 2020