Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

DenseSR: Image Shadow Removal as Dense Prediction

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read Single-image shadow removal is treated as a dense prediction problem, with the paper claiming that combining deep geometric-semantic scene priors with an adaptive dense-fusion decoder achieves state-of-the-art mask-free shadow removal.

desk verdict Solid incremental shadow-removal architecture, but the SOTA claim is contradicted by its own Table 1; worth peer review after fixes. read the letter →

arxiv 2507.16472 v1 pith:J7NIMN2O submitted 2025-07-22 cs.CV

classification cs.CV
keywords single-imageshadowremovaldensepredictionscenepriorsgeometric-semanticguidanceadaptivefeaturefusioncontentsmoothingboundaryrestorationmask-free
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

Single-image shadow removal is treated as a dense prediction problem, and the paper's central claim is that combining deep geometric-semantic scene priors with an adaptive dense-fusion decoder produces state-of-the-art mask-free shadow removal. The paper argues that shadows non-uniformly degrade multi-scale features, so standard fusion blurs boundaries and leaves intra-shadow inconsistency; DenseSR instead separates content smoothing from detail recuperation before recombination. The model fuses frozen depth, surface-normal, and DINO semantic features through attention modulation to resolve ambiguity between shadows and intrinsically dark surfaces, then restores the image with the Dense Fusion Block. On ISTD, ISTD+, SRD, WSRD+, and INS benchmarks it reports the best or near-best PSNR/SSIM among mask-free methods, and remains competitive with methods that use ground-truth masks. If right, the work establishes that implicit shadow localization via scene priors plus component-aware fusion is a practical alternative to mask-guided and diffusion-based restoration.

What carries the argument

The Dense Fusion Block (DFB) is the central mechanism: at each decoder upsampling stage it splits the incoming features into low-frequency and high-frequency components, applies adaptive content-aware kernels (ACSM with softmax-normalized low-pass kernels for intra-shadow consistency, and TBRM with inverted low-pass kernels as high-pass filters for texture and boundary recovery), and recombines them via CARAFE reassembly and element-wise addition. The companion mechanism is the Scene-Integrated Module (SIM), which makes window self-attention explicitly depend on scene priors by computing pairwise semantic similarity from DINO feature dot products and geometric consistency from planar distances derived with depth and normals, and multiplying those maps into the QK^T scores before softmax.

What would settle it

Run the paper's ablations with the normal-estimation field of view changed to, say, 45 and 90 degrees, or with the depth and DINO features deliberately corrupted while keeping the network fixed: if PSNR/SSIM on ISTD+ or INS does not drop beyond the reported margins, the scene-prior pathway is not load-bearing. A simpler check is an independent per-image significance test on the Table 3 differences: if removing depth or normals never changes results outside the evaluation noise floor, the geometric-prior claim collapses.

Watch

Extended reading notes

Core claim

The paper's discovery is that the main bottleneck in modern shadow removal is not just scene understanding but the restoration-stage feature fusion itself: shadows degrade deep features non-uniformly, so the decoder has to process content and detail components separately before combining them. DenseSR's Dense Fusion Block predicts spatially-variant low-pass kernels to smooth the base appearance inside shadows for consistency and spatially-variant high-pass kernels to re-inject textures and sharpen boundaries, then fuses the two streams after content-aware reassembly. Scene-Integrated Modules in the encoder and bottleneck modulate window attention with DINO-derived semantic similarity maps and depth/normal-derived geometric consistency maps, which lets the network implicitly localize shadows without a mask. With this design, the reported numbers on ISTD, ISTD+, SRD, WSRD+, and INS are the best or second-best among mask-free methods, and the ablations show each prior and each DFB sub-module contributes to the final score.

Load-bearing premise

The load-bearing premise is that the frozen depth and DINO features, plus the assumed 60-degree field of view for normal estimation, give trustworthy geometric and semantic cues for shadow disambiguation; if those priors are noisy or the camera intrinsics differ, the attention modulation can mislead restoration.

Editorial extensions

If this is right

  • Mask-free shadow removal can reach or approach mask-guided quality on standard benchmarks, so real applications no longer need an accurate shadow mask at inference time.
  • Frozen geometric and semantic priors reduce shadow ambiguity enough to substitute for explicit shadow detection in the restoration pipeline.
  • Splitting restoration into smoothed base content and reinjected high-frequency detail improves both intra-shadow consistency and boundary sharpness relative to standard feature fusion, as the ablations and optimization curves show.
  • DenseSR runs in about 125 ms on a 640x480 image, faster than diffusion-based shadow removal, while matching or beating those methods on quality.
  • The decoder's frequency spectra show stronger high-frequency energy at each stage when DFB is used, which the paper connects to sharper boundaries and finer textures.

Reading between the lines

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

  • Because the priors are frozen and the attention modulation is generic, DenseSR's architecture could transfer to other spatially non-uniform restoration tasks such as low-light enhancement, reflection removal, or dehazing; the paper does not test this.
  • The normal map depends on a fixed 60-degree field-of-view assumption, so on images with different camera intrinsics the normals will be biased; testing with calibrated intrinsics or omitting normals would reveal whether this prior is robust or dataset-specific.
  • A per-image breakdown of the reported averages would test whether the gains concentrate on the indoor, indirect-lighting images the paper emphasizes, or spread uniformly across all benchmarks.
  • Fine-tuning or adaptively weighting the depth and DINO priors instead of freezing them could push performance further, at greater training cost; the paper leaves this unexplored.
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.

Referee Report

4 major / 5 minor

Summary. The paper proposes DenseSR, a mask-free single-image shadow removal network. It augments a U-Net-style Swin Transformer encoder with depth, surface-normal, and DINO-V2 semantic priors via a Scene-Integrated Module (SIM), and inserts a Dense Fusion Block (DFB) in the decoder. The DFB splits restoration into an Adaptive Content Smoothing Module (ACSM) for low-frequency consistency and a Texture-Boundary Recuperation Module (TBRM) for high-frequency detail, fusing the two paths with CARAFE. The model is trained with Charbonnier loss and evaluated on ISTD, ISTD+, SRD, WSRD+, and INS. The paper claims state-of-the-art performance in mask-free shadow removal.

Significance. If the reported numbers are reliable, DenseSR advances a timely direction: combining frozen geometric and semantic priors with adaptive feature fusion for shadow removal. The component ablations in Table 3 give evidence that depth, normals, DINO features, and the DFB each help, and the method is substantially faster than diffusion baselines. However, the manuscript's headline claim is currently overstated: Table 1 shows that StableShadowDiffusion beats DenseSR on ISTD+ and SRD, and the margins over OmniSR are small and unreplicated. Because no error bars, multiple seeds, or released code are provided, the SOTA claim is not yet established. The work is best assessed as a solid architectural contribution whose headline claim needs to be either statistically supported or scaled back.

major comments (4)
  1. [Sec. 5.2, Table 1] The unqualified 'state-of-the-art' claim in the abstract and conclusion is not supported by Table 1. In the mask-free rows, StableShadowDiffusion [60] outperforms DenseSR by 1.21 dB on ISTD+ (35.19 vs. 33.98) and by 0.18 dB on SRD (33.63 vs. 33.45); on WSRD+ the lead over StableShadowDiffusion is 0.02 dB, and the gains over OmniSR are 0.19, 0.64, 0.58, and 0.21 dB across the four datasets. Since no error bars or repeated-seed results are given, these margins may lie within run-to-run variation. The authors should either soften the SOTA wording (as the 'competitive' phrasing in Sec. 5.2 already does) or provide a proper statistical comparison.
  2. [Sec. 5.2, +GM paragraph] The text says DenseSR+GM is 'second-best' on ISTD, 'surpassed only by ShadowDiffusion + GM and ShadowFormer + GM,' but Table 1 lists DenseSR+GM at 32.14 PSNR, ShadowFormer+GM at 32.21, and ShadowDiffusion+GM at 32.33; it is therefore third, not second. This internal inconsistency should be corrected.
  3. [Supplementary Eq. (8), Sec. 4.1.1] The normal map is derived from depth using a fixed 60-degree field of view (Eq. 8 in the supplementary). The benchmark datasets contain images from different cameras, and the paper gives no justification for this value or any sensitivity analysis. Because Table 3 shows that removing the normal prior costs 1.32 dB on INS testing and 0.48 dB on WSRD+, an incorrect FOV could materially bias an input that the method depends on. The authors should validate the FOV choice or evaluate robustness to this hyperparameter.
  4. [Table 3] The individual ablations of ACSM and TBRM are small: removing ACSM costs 0.12 dB (INS testing) and 0.17 dB (WSRD+), and removing TBRM costs 0.15 dB and 0.21 dB. Without error bars, multiple seeds, or paired significance tests, these differences are not distinguishable from stochastic variation. The claim that both modules demonstrably contribute needs statistical support.
minor comments (5)
  1. [Abstract and Sec. 6] The abstract and conclusion use 'state-of-the-art' while Sec. 5.2 correctly says 'competitive PSNR and SSIM scores'; unify the wording after addressing the corresponding major comment.
  2. [Abstract] The abstract says 'Our code can be available' but no release is indicated; please state clearly whether and where the code will be released.
  3. [Sec. 5.2 and Figure 6] The method of reference [60] is called 'StableShadowRefiner' in one place and 'StableShadowDiffusion' elsewhere; use one consistent name.
  4. [Table 1] The caption says the best results are highlighted as 1st, 2nd, and 3rd, but the table as rendered has no visible highlighting; clarify or add the formatting.
  5. [Sec. 5.2] The text mentions evaluation on INS alongside ISTD, ISTD+, SRD, and WSRD+, but INS appears only in Table 2; add a cross-reference so the reader can locate the INS numbers.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: DenseSR's contribution is evaluated against external benchmarks; the only load-bearing issue is the SOTA claim contradicted by its own Table 1, which is a correctness concern, not circularity.

full rationale

The paper's central claim is empirical: DenseSR achieves strong shadow-removal quality on ISTD, ISTD+, SRD, WSRD+, and INS, as measured by PSNR/SSIM against external ground truth in Tables 1 and 2. There is no derivation that reduces to its own inputs. The model is trained with the Charbonnier loss in Eq. (7) supervised by ground-truth shadow-free images, and the ablations in Table 3 are retrained on the same external benchmarks, so the contribution of each module is measured rather than constructed. The normal-map computation in Eqs. (8)-(11) of the supplementary is a standard pinhole back-projection with a stated 60-degree FOV assumption; it is an external geometric conversion, not a fit to the shadow-removal target. The self-citations present (e.g., DRCT as reference [19], PromptHSI as reference [27]) are used only as general related-work motivation or as prior examples of using pre-trained priors, not as the justification for DenseSR's architecture or results. No uniqueness theorem, no ansatz smuggled via citation, and no fitted parameter renamed as a prediction were found. The paper does have a substantive internal-consistency problem: its own Table 1 shows StableShadowDiffusion achieving 35.19 PSNR on ISTD+ and 33.63 on SRD versus DenseSR's 33.98 and 33.45, so the abstract's 'state-of-the-art' wording is overstated, and the Section 5.2 text describing DenseSR+GM as 'surpassed only by' two methods is contradicted by the table. These are correctness and evidence-quality concerns, however, not circularity; the evaluation remains self-contained against external benchmarks. The appropriate circularity score is therefore low.

Assumptions & free parameters 5 free parameters · 4 assumptions · 3 invented entities

The paper is an empirical architecture paper. The listed free parameters are standard training and module hyperparameters chosen by the authors; they are not derived from theory. The axioms are the domain assumptions that any supervised learning method of this type makes. The invented entities are new network components rather than physical entities, and they are validated only through the paper's own experiments.

free parameters (5)
  • FOV for normal estimation = 60 degrees
    Chosen in supplementary Eq. 8; affects all normal maps computed from depth, and thus the geometric consistency maps in SIM attention.
  • ACSM low-pass kernel size K = 5
    Set in supplementary Sec. 7.2; determines smoothing neighborhood in the content consistency module.
  • TBRM high-pass kernel size = 3
    Set in supplementary Sec. 7.2; controls locality of detail extraction in the texture restoration module.
  • Base embedding dimension = 32
    Network width; changing it would alter capacity and likely performance.
  • Charbonnier loss epsilon = 1e-3
    Numerical stability constant in Eq. 7.
assumptions (4)
  • domain assumption Shadowed image approximately equals shadow-free image times a spatially varying illumination factor A(x)
    Stated in Sec 3.1 as a simplification; the paper acknowledges non-linear effects, but the multiplicative model motivates the training setup.
  • domain assumption DepthAnything-V2 and DINO-V2 priors are reliable enough to improve attention
    Used in Sec 4.1.1 and 4.1.2; no failure-case analysis is provided for inaccurate or noisy priors.
  • domain assumption Ground-truth shadow-free images in ISTD, ISTD+, SRD, WSRD+, and INS are accurate targets
    The loss in Eq. 7 assumes these ground truths are correct; no uncertainty modeling or label noise handling is described.
  • standard math Softmax-normalized filter weights plus identity subtraction yields valid high-pass filters
    Filter inversion in Eq. 4 relies on this signal-processing identity, which is standard but made data-dependent through the softmax.
invented entities (3)
  • Adaptive Content Smoothing Module (ACSM)
    purpose: Predicts spatially-varying smoothing kernels to regularize content consistency in shadow regions.
    Proposed in this paper; only evidence is the benchmark ablation in Table 3.
  • Texture-Boundary Recuperation Module (TBRM)
    purpose: Predicts high-pass kernels to recover textures and sharpen shadow boundaries.
    Proposed in this paper; no external validation beyond the paper's experiments.
  • Dense Fusion Block (DFB)
    purpose: Combines ACSM and TBRM outputs with CARAFE to fuse cross-scale features in the decoder.
    The core architectural contribution; no independent evidence outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DenseSR: Image Shadow Removal as Dense Prediction." pith.science (2026). https://pith.science/paper/J7NIMN2O

@misc{pith2026250716472,
  author       = {Pith},
  title        = {Pith review of: DenseSR: Image Shadow Removal as Dense Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J7NIMN2O}},
  note         = {Machine review of arXiv:2507.16472}
}
abstract

Shadows are a common factor degrading image quality. Single-image shadow removal (SR), particularly under challenging indirect illumination, is hampered by non-uniform content degradation and inherent ambiguity. Consequently, traditional methods often fail to simultaneously recover intra-shadow details and maintain sharp boundaries, resulting in inconsistent restoration and blurring that negatively affect both downstream applications and the overall viewing experience. To overcome these limitations, we propose the DenseSR, approaching the problem from a dense prediction perspective to emphasize restoration quality. This framework uniquely synergizes two key strategies: (1) deep scene understanding guided by geometric-semantic priors to resolve ambiguity and implicitly localize shadows, and (2) high-fidelity restoration via a novel Dense Fusion Block (DFB) in the decoder. The DFB employs adaptive component processing-using an Adaptive Content Smoothing Module (ACSM) for consistent appearance and a Texture-Boundary Recuperation Module (TBRM) for fine textures and sharp boundaries-thereby directly tackling the inconsistent restoration and blurring issues. These purposefully processed components are effectively fused, yielding an optimized feature representation preserving both consistency and fidelity. Extensive experimental results demonstrate the merits of our approach over existing methods. Our code can be available on https://github$.$com/VanLinLin/DenseSR

Figures

Figures reproduced from arXiv: 2507.16472 by the authors.

Figure 1
Figure 1. To tackle inconsistent restoration and boundary [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Evolution of shadow removal approaches. (a) Mask [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overall architecture of the proposed DenseSR framework, embodying a layered ’understand then restore’ strategy. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Multi-modal scene priors and feature map within [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: The SIM leverages geometric and semantic priors to make its self-attention context-aware, enhancing scene under [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Results visualization with SOTA shadow removal methods, illustrating reconstruction quality of our method on [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 8
Figure 8. Figure 8: Optimization curve between with/without the pro [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 7
Figure 7. Figure 7: Real-world testing data comparisons. For the real [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Sign in 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

  1. Anchoring Trends: Mitigating Social Media Popularity Prediction Drift via Feature Clustering and Expansion

    cs.MM 2025-07 reject novelty 5.0 of 10

    A multimodal clustering and LLM feature generation framework for social media popularity prediction is proposed, but its temporal robustness claim is not supported by the evaluation design.

Reference graph

Works this paper leans on

71 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [60]

    Jiamin Xu, Yuxin Zheng, Zelong Li, Chi Wang, Renshu Gu, Weiwei Xu, and Gang Xu. 2024. Detail-Preserving Latent Diffusion for Stable Shadow Removal. arXiv:2412.17630 [cs.CV] https://arxiv.org/abs/2412.17630

  2. [1]

    R. B. Blackman and J. W. Tukey. 1958. The measurement of power spectra from the point of view of communications engineering — Part I. The Bell System Technical Journal 37, 1 (1958), 185–282. doi:10.1002/j.1538-7305.1958.tb03874.x

  3. [2]

    Luis Bolanos, Shih-Yang Su, and Helge Rhodin. 2024. Gaussian Shadow Casting for Neural Characters. In The Conference on Computer Vision and Pattern Recognition

  4. [3]

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexan- der Kirillov, and Sergey Zagoruyko. 2020. End-to-end object detection with transformers. In European conference on computer vision . Springer, 213–229

  5. [4]

    Rita Cucchiara, Costantino Grana, Massimo Piccardi, and Andrea Prati. 2003. Detecting moving objects, ghosts, and shadows in video streams.IEEE transactions on pattern analysis and machine intelligence 25, 10 (2003), 1337–1342

  6. [5]

    Xiaodong Cun, Chi-Man Pun, and Cheng Shi. 2020. Towards Ghost-Free Shadow Removal via Dual Hierarchical Aggregation Network and Shadow Matting GAN. 10680–10687

  7. [6]

    Daboczi and T.B

    T. Daboczi and T.B. Bako. 2000. Inverse filtering of optical images. In Proceedings of the 17th IEEE Instrumentation and Measurement Technology Conference [Cat. No. 00CH37066], Vol. 1. 370–374 vol.1. doi:10.1109/IMTC.2000.846888

  8. [7]

    Rakesh Kumar Das, Madhu Shandilya, Shubham Sharma, and Dhanshree Kulkarni

Show all 71 references
  1. [9]

    Wei Dong, Han Zhou, Yuqiong Tian, Jingke Sun, Xiaohong Liu, Guangtao Zhai, and Jun Chen. 2024. ShadowRefiner: Towards Mask-free Shadow Removal via Fast Fourier Transformer. arXiv:2406.02559 [cs.CV] https://arxiv.org/abs/2406.02559

  2. [10]

    Lan Fu, Changqing Zhou, Qing Guo, Felix Juefei-Xu, Hongkai Yu, Wei Feng, Yang Liu, and Song Wang. 2021. Auto-exposure fusion for single-image shadow removal. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 10571–10580

  3. [11]

    A Ardeshir Goshtasby. 2005. Fusion of multi-exposure images. Image and Vision Computing 23, 6 (2005), 611–618

  4. [12]

    Lanqing Guo, Siyu Huang, Ding Liu, Hao Cheng, and Bihan Wen. 2023. Shad- owformer: global context helps shadow removal. In Proceedings of the AAAI conference on artificial intelligence , Vol. 37. 710–718

  5. [13]

    Laniqng Guo, Chong Wang, Yufei Wang, Siyu Huang, Wenhan Yang, Alex C Kot, and Bihan Wen. 2024. Single-Image Shadow Removal Using Deep Learning: A Comprehensive Survey. arXiv preprint arXiv:2407.08865 (2024)

  6. [14]

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

  7. [15]

    Bae Gwangbin, Budvytis Ignas, and Cipolla Roberto. 2021. Estimating and Ex- ploiting the Aleatoric Uncertainty in Surface Normal Estimation. In International Conference on Computer Vision (ICCV)

  8. [16]

    Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. 2017. Mask R- CNN. In 2017 IEEE International Conference on Computer Vision (ICCV). 2980–2988. doi:10.1109/ICCV.2017.322

  9. [17]

    Shengfeng He, Bing Peng, Junyu Dong, and Yong Du. 2021. Mask-ShadowNet: Toward Shadow Removal via Masked Adaptive Instance Normalization. IEEE Signal Processing Letters 28 (2021), 957–961. doi:10.1109/LSP.2021.3074082

  10. [18]

    Chih-Chung Hsu, Chih-Yu Jian, Eng-Shen Tu, Chia-Ming Lee, and Guan-Lin Chen

  11. [19]

    Chih-Chung Hsu, Chia-Ming Lee, and Yi-Shiuan Chou. 2024. DRCT: Saving Image Super-Resolution Away from Information Bottleneck. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. 6133–6142

  12. [20]

    Chih-Chung Hsu, Chia-Ming Lee, Yu-Fan Lin, Yi-Shiuan Chou, Chih-Yu Jian, and Chi-Han Tsai. 2024. Revisiting Vision-Language Features Adaptation and Inconsistency for Social Media Popularity Prediction. In Proceedings of the 32nd ACM International Conference on Multimedia (Melb...

  13. [21]

    Xiaowei Hu, Chi-Wing Fu, Lei Zhu, Jing Qin, and Pheng-Ann Heng. 2019. Direction-aware spatial context features for shadow detection and removal. IEEE transactions on pattern analysis and machine intelligence 42, 11 (2019), 2795–2808

  14. [22]

    Huang, and Humphrey Shi

    Zilong Huang, Yunchao Wei, Xinggang Wang, Wenyu Liu, Thomas S. Huang, and Humphrey Shi. 2022. AlignSeg: Feature-Aligned Segmentation Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence 44, 1 (2022), 550–557. doi:10.1109/TPAMI.2021.3062772

  15. [23]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. 2015. Adam: A Method for Stochastic Opti- mization. In ICLR

  16. [24]

    Hieu Le and Dimitris Samaras. 2019. Shadow removal via shadow image decom- position. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 8578–8587

  17. [25]

    Hieu Le and Dimitris Samaras. 2019. Shadow Removal via Shadow Image Decom- position. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)

  18. [26]

    Hieu Le and Dimitris Samaras. 2020. From Shadow Segmentation to Shadow Removal

  19. [27]

    Chia-Ming Lee, Ching-Heng Cheng, Yu-Fan Lin, Yi-Ching Cheng, Wo-Ting Liao, Fu-En Yang, Yu-Chiang Frank Wang, and Chih-Chung Hsu. 2025. PromptHSI: Universal Hyperspectral Image Restoration with Vision-Language Modulated Frequency Adaptation. arXiv:2411.15922 [eess.IV] https://a...

  20. [28]

    Youngwan Lee, Jonghee Kim, Jeffrey Willette, and Sung Ju Hwang. 2022. MPViT: Multi-Path Vision Transformer for Dense Prediction. In2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . 7277–7286. doi:10.1109/ CVPR52688.2022.00714

  21. [29]

    Chenghua Li, Bo Yang, Zhiqi Wu, Gao Chen, Yihan Yu, and Shengxiao Zhou

  22. [30]

    Tsung-Yi Lin, Piotr Dollár, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. 2017. Feature Pyramid Networks for Object Detection. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) . 936–944. doi:10.1109/CVPR.2017.106

  23. [31]

    Hengxing Liu, Mingjia Li, and Xiaojie Guo. 2024. Regional Attention for Shadow Removal. arXiv:2411.14201 [cs.CV] https://arxiv.org/abs/2411.14201

  24. [33]

    Jiawei Liu, Qiang Wang, Huijie Fan, Jiandong Tian, and Yandong Tang. 2023. A Shadow Imaging Bilinear Model and Three-Branch Residual Network for Shadow Removal. IEEE Transactions on Neural Networks and Learning Systems (2023), 1–15. doi:10.1109/TNNLS.2023.3290078

  25. [34]

    Wenze Liu, Hao Lu, Hongtao Fu, and Zhiguo Cao. 2023. Learning to Upsample by Learning to Sample. In 2023 IEEE/CVF International Conference on Computer Vision (ICCV). 6004–6014. doi:10.1109/ICCV51070.2023.00554

  26. [35]

    Jiawei Liu, Qiang Wang, Huijie Fan, Wentao Li, Liangqiong Qu, and Yandong Tang

  27. [36]

    Hao Lu, Wenze Liu, Zixuan Ye, Hongtao Fu, Yuliang Liu, and Zhiguo Cao. 2022. SAPA: Similarity-Aware Point Affiliation for Feature Upsampling. InProc. Annual Conference on Neural Information Processing Systems (NeurIPS)

  28. [37]

    Kangfu Mei, Luis Figueroa, Zhe Lin, Zhihong Ding, Scott Cohen, and Vishal M Patel. 2024. Latent Feature-Guided Diffusion Models for Shadow Removal. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision . 4313–4322

  29. [38]

    Saritha Murali, VK Govindan, and Saidalavi Kalady. 2016. A survey on shadow removal techniques for single image. International Journal of Image, Graphics and Signal Processing 8, 12 (2016), 38

  30. [39]

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. 2021. Swin transformer: Hierarchical vision transformer us- ing shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision. 10012–10022

  31. [40]

    and Piotr Bojanowski

    Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mahmoud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michael Rabba...

  32. [41]

    Liangqiong Qu, Jiandong Tian, Shengfeng He, Yandong Tang, and Rynson WH Lau. 2017. Deshadownet: A multi-context embedding deep network for shadow removal. 4067–4075

  33. [42]

    René Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. 2021. Vision Transformers for Dense Prediction. In 2021 IEEE/CVF International Conference on Computer Vision (ICCV). 12159–12168. doi:10.1109/ICCV48922.2021.01196

  34. [43]

    Kunpeng Niu, Yanli Liu, Enhua Wu, and Guanyu Xing. 2023. A Boundary-Aware Network for Shadow Removal. IEEE Transactions on Multimedia 25 (2023), 6782–

  35. [44]

    Neda Salamati, Arthur Germain, and S Siisstrunk. 2011. Removing shadows from images using color and near-infrared. In 2011 18th IEEE International Conference on Image Processing. IEEE, 1713–1716

  36. [45]

    Andres Sanin, Conrad Sanderson, and Brian C. Lovell. 2010. Improved Shadow Removal for Robust Person Tracking in Surveillance Scenarios. In 2010 20th International Conference on Pattern Recognition . 141–144. doi:10.1109/ICPR.2010. 43

  37. [46]

    Prafull Sharma, Julien Philip, Michaël Gharbi, Bill Freeman, Fredo Durand, and Valentin Deschaintre. 2023. Materialistic: Selecting Similar Materials in Images. ACM Trans. Graph. 42, 4, Article 154 (jul 2023), 14 pages. doi:10.1145/3592390

  38. [47]

    Evan Shelhamer, Jonathan Long, and Trevor Darrell. 2017. Fully Convolutional Networks for Semantic Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence 39, 4 (2017), 640–651. doi:10.1109/TPAMI.2016.2572683

  39. [48]

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. 2015. U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical image computing and computer-assisted intervention . Springer, 234–241

  40. [49]

    Yael Shor and Dani Lischinski. 2008. The shadow meets the mask: Pyramid-based shadow removal. In Computer Graphics Forum , Vol. 27. Wiley Online Library, 577–586

  41. [50]

    Arti Tiwari, Pradeep Kumar Singh, and Sobia Amin. 2016. A survey on shadow detection and removal in images and video sequences. In 2016 6th International Conference-Cloud System and Big Data Engineering (Confluence) . IEEE, 518–523

  42. [51]

    Florin-Alexandru Vasluianu, Tim Seizinger, and Radu Timofte. 2023. WSRD: A Novel Benchmark for High Resolution Image Shadow Removal. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) . 1826–1835. doi:10.1109/CVPRW59228.2023.00181

  43. [52]

    Florin-Alexandru Vasluianu, Tim Seizinger, Zhuyun Zhou, Zongwei Wu, Cailian Chen, and Radu Timofte. 2024. NTIRE 2024 Image Shadow Removal Challenge Report. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). 6547–6570. doi:10.1109/CVPRW633...

  44. [53]

    Aitken, Rob Bishop, Daniel Rueckert, and Zehan Wang

    Wenzhe Shi, Jose Caballero, Ferenc Huszár, Johannes Totz, Andrew P. Aitken, Rob Bishop, Daniel Rueckert, and Zehan Wang. 2016. Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network. In 2016 IEEE Conference on Computer Visio...

  45. [54]

    Jifeng Wang, Xiang Li, and Jian Yang. 2018. Stacked conditional generative adversarial networks for jointly learning shadow detection and shadow removal. 1788–1797

  46. [55]

    Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and Ling Shao. 2021. Pyramid vision transformer: A versatile back- bone for dense prediction without convolutions. In Proceedings of the IEEE/CVF International Conference on Computer Vi...

  47. [56]

    Zhendong Wang, Xiaodong Cun, Jianmin Bao, Wengang Zhou, Jianzhuang Liu, and Houqiang Li. 2022. Uformer: A general u-shaped transformer for image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 17683–17693

  48. [57]

    Brostow, Michael Firman, and Sara Vicente

    Silvan Weder, Guillermo Garcia-Hernando, Áron Monszpart, Marc Pollefeys, Gabriel J. Brostow, Michael Firman, and Sara Vicente. 2023. Removing Objects From Neural Radiance Fields. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR) . 16...

  49. [58]

    Jiaqi Wang, Kai Chen, Rui Xu, Ziwei Liu, Chen Change Loy, and Dahua Lin. 2019. CARAFE: Content-Aware ReAssembly of FEatures. In The IEEE International Conference on Computer Vision (ICCV)

  50. [59]

    Jiamin Xu, Zelong Li, Yuxin Zheng, Chenyu Huang, Renshu Gu, Weiwei Xu, and Gang Xu. 2025. OmniSR: Shadow Removal under Direct and Indirect Lighting. arXiv:2410.01719 [cs.CV] https://arxiv.org/abs/2410.01719

  51. [61]

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. 2024. Depth Anything V2. arXiv:2406.09414 [cs.CV] https://arxiv.org/abs/2406.09414

  52. [62]

    Qingxiong Yang, Kar-Han Tan, and Narendra Ahuja. 2012. Shadow removal using bilateral filtering. IEEE Transactions on Image processing 21, 10 (2012), 4361–4368

  53. [63]

    Jie Xiao, Xueyang Fu, Yurui Zhu, Dong Li, Jie Huang, Kai Zhu, and Zheng-Jun Zha

  54. [64]

    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    HomoFormer: Homogenized Transformer for Image Shadow Removal. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 25617–25626

  55. [65]

    Yurui Zhu, Jie Huang, Xueyang Fu, Feng Zhao, Qibin Sun, and Zheng-Jun Zha

  56. [66]

    Yurui Zhu, Zeyu Xiao, Yanchi Fang, Xueyang Fu, Zhiwei Xiong, and Zheng-Jun Zha. 2022. Efficient model-driven network for shadow removal. In Proceedings of the AAAI conference on artificial intelligence , Vol. 36. 3635–3643. DenseSR: Image Shadow Removal as Dense Prediction MM’...

  57. [69]

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, and Ling Shao. 2020. Learning enriched features for real image restoration and enhancement. In IEEE/CVF European Conference on Computer Vision (ECCV)

  58. [70]

    Ling Zhang, Qing Zhang, and Chunxia Xiao. 2015. Shadow remover: Image shadow removal based on illumination recovering optimization.IEEE Transactions on Image Processing 24, 11 (2015), 4623–4636

  59. [2017]

    In 2017 International Conference on Recent Innovations in Signal processing and Embedded Systems (RISE)

    A survey on shadow detection and removal in images. In 2017 International Conference on Recent Innovations in Signal processing and Embedded Systems (RISE). 175–180. doi:10.1109/RISE.2017.8378149

  60. [2022]

    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Bijective Mapping Network for Shadow Removal. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 5627–5636

  61. [2023]

    IEEE Transactions on Multimedia (2023)

    A Decoupled Multi-Task Network for Shadow Removal. IEEE Transactions on Multimedia (2023)

  62. [2024]

    IEEE Transactions on Geoscience and Remote Sensing 62 (2024), 1–16

    Real-Time Compressed Sensing for Joint Hyperspectral Image Transmission and Restoration for CubeSat. IEEE Transactions on Geoscience and Remote Sensing 62 (2024), 1–16. doi:10.1109/TGRS.2024.3378828

  63. [6793]

    doi:10.1109/TMM.2022.3214422

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.