Pith. sign in

REVIEW 3 major objections 6 minor 124 references

Adapting a video diffusion model with LoRA and shadow-mask-guided detail injection yields temporally consistent shadow removal in unconstrained videos.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 00:29 UTC pith:ZZPIL4UB

load-bearing objection First video-diffusion video shadow removal with a substantial synthetic dataset, but the in-the-wild claim outruns the evidence: real-world evaluation is qualitative only. the 3 major comments →

arxiv 2607.26203 v1 pith:ZZPIL4UB submitted 2026-07-28 cs.CV

WildShadowRemover: In-the-Wild Video Shadow Removal via Detail-Preserving Video Diffusion Models

Jiamin Xu , Cong Wang , Zheng Dong , Chi Wang , Renshu Gu , Weiwei Xu , Gang Xu This is my paper
classification cs.CV
keywords video shadow removalvideo diffusion modelsLoRA fine-tuningdetail preservationtemporal consistencysynthetic datasetshadow maskin-the-wild generalization
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Existing video shadow removal methods work frame-by-frame and lose temporal coherence. This paper argues that a pretrained video diffusion model carries enough implicit knowledge of how shadows move and change with objects that only a light-weight LoRA fine-tuning step is needed to turn it into a video shadow remover. To stop the diffusion model's latent decoding from blurring away fine texture, the authors insert a shadow-mask-guided frequency-decomposed modulation into the frozen VAE decoder, which restores high-frequency detail where the mask signals shadow while keeping shadow patterns out. They also condition on monocular depth to disambiguate shadows under complex lighting, and train on WildShadow, a large new synthetic dataset spanning indoor, urban, and natural scenes. The paper's claim is that the resulting model beats per-frame image methods on reconstruction quality and temporal consistency and generalizes to real videos, with real-world evaluation shown qualitatively.

Core claim

On the paper's own terms, the central discovery is that a video foundation model's spatiotemporal prior, not a task-specific temporal architecture, is what solves video shadow removal. The method performs shadow removal in the latent space of a pretrained video diffusion model: a diffusion transformer conditioned on the shadow video and a depth video is fine-tuned with LoRA using a flow-matching objective, producing a coarse shadow-free latent that is decoded to a coarse video. Because the frozen decoder loses thin structures and textures, the authors augment each decoder stage with a detail injection module driven by a soft shadow mask computed as the ratio between input and coarse output.

What carries the argument

The carrying mechanism is a two-stage pipeline built around a pretrained latent video diffusion model. Stage one is low-rank (LoRA) fine-tuning of the diffusion transformer, conditioned on the input shadow video and on monocular depth maps, trained with a flow-matching objective to predict a coarse shadow-free latent video. Stage two is a shadow-mask-guided detail injection module: a soft shadow mask M = 1 − X/(Y0 + ε) is estimated by comparing the input with the coarse prediction, and at each decoder stage the encoder and decoder features are concatenated, split into low- and high-frequency components using a fixed 5×5 mean filter, and independently modulated by mask-predicted affine parame

Load-bearing premise

The load-bearing premise is that the synthetic dataset, whose shadow-free targets are made by removing only direct-light shadows in rendered scenes, is representative enough of real in-the-wild shadow videos for the trained model to generalize — a premise the paper tests only qualitatively on real footage.

What would settle it

Record the same real scene twice with a movable light or occluder so that one video has shadows and the other is shadow-free but otherwise identical; run the trained model and a strong frame-by-frame image method on the shadowed video, and compare both against the physically truthful shadow-free frames using PSNR/SSIM plus a temporal-consistency metric. If the video model does not beat the per-frame baseline on temporal consistency and quality, the central claim fails.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If the method is right, video shadow removal no longer needs per-frame post-processing to stay temporally stable; the diffusion model supplies coherence across frames.
  • Clean shadow-free video can be piped into downstream tasks such as video editing, object tracking, intrinsic decomposition, and 3D content creation, which have all been hampered by shadow artifacts.
  • The new large paired synthetic dataset, with its benchmark protocol, gives the community a common training and evaluation resource for video shadow removal.
  • Because the same model handles single images, a single video diffusion-based remover could unify image and video shadow removal, matching specialized image methods on standard benchmarks.

Where Pith is reading between the lines

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

  • The real-world 'in-the-wild' claim rests only on qualitative examples; a quantitative evaluation on real videos with physically captured ground truth would be the natural test and is not in the paper.
  • The frequency-decomposed modulation idea could be extracted as a general detail-restoration module for other conditional diffusion models where residual artifacts and texture are entangled.
  • Given that depth conditioning adds only a small improvement in the ablations, a cheaper or more robust geometry cue may capture the same benefit; this is an inexpensive experiment to run.
  • The dependence on synthetic data means the method's ceiling is set by how faithfully renderers model real illumination; self-supervised adaptation on real videos is the stated open problem.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes WildShadowRemover, a video shadow removal framework built on a pretrained video diffusion model (Wan-Control). It adapts the DiT through LoRA fine-tuning, condenses the input shadow video and Depth Anything 3 depth maps, and augments the frozen VAE decoder with a shadow-mask-guided detail injection module that uses frequency-decomposed modulation to restore high-frequency details without re-injecting shadow remnants. The authors also introduce WildShadow, a large-scale synthetic video dataset rendered from 3D-FRONT, iCity, and Infinigen, with paired shadow/shadow-free clips, and a related image dataset (WildShadow-I) for cross-domain evaluation. Experiments compare against image shadow removal baselines on ISTD+, SRD, INS, and WildShadow-I, and include an ablation study, a synthetic video benchmark comparison (against OmniSR and a frame-by-frame variant), and qualitative examples on real DL3DV and UAV videos. The paper claims state-of-the-art performance and strong generalization to in-the-wild videos.

Significance. If the claims were fully supported, this would be among the first video-diffusion-based video shadow removal methods and would provide a useful large-scale synthetic benchmark for the community. The technical components are reasonable and mostly standard: flow matching, LoRA adaptation, depth conditioning, and a residual detail-injection module; the ablations in Table 3 give independent evidence for each component, and the paper does not rely on circular or parameter-fitting claims. The main weakness is that the central 'in-the-wild generalization' claim is supported only by qualitative examples on real videos, while all quantitative evidence is on synthetic benchmarks or image datasets. Additionally, the only prior video shadow removal method (PSTNet) is cited but never compared, leaving the 'state-of-the-art' video claim unsubstantiated. The method is not presented as SOTA on conventional image benchmarks, and the abstract overstates the results relative to Table 1. The dataset contribution is potentially valuable, though reproducibility would benefit from a clear release/access statement.

major comments (3)
  1. [Real-World Evaluation; Limitations] The core claim of strong generalization to in-the-wild videos is not supported by the evidence shown. Real-world evaluation on DL3DV and UAV videos (Fig. 5) is purely qualitative, with no ground truth, no metrics, no user study, and no temporal-consistency measurement. The only quantitative evidence (Tables 1, 2) comes from synthetic benchmarks, while the Limitations section explicitly states that the framework 'relies primarily on synthetic paired data for training' and that leveraging real-world videos remains open. Moreover, the dataset section says outdoor shadow-free targets remove only shadows cast by direct illumination, whereas real videos contain penumbrae, interreflections, and ambient occlusion; the size of this domain gap is never measured. Please add quantitative real-world evaluation (e.g., a paired real-video subset, no-reference or perceptual metrics, temporal-consistency
  2. [Video Shadow Removal Evaluation; Related Work] PSTNet (Chen et al. 2024) is the only prior video shadow removal method cited, but it is never compared against. Table 4 compares only OmniSR (image-based, frame-by-frame) and the authors' own frame-by-frame variant. Without a PSTNet comparison under the same protocol, the claim of 'state-of-the-art performance' and 'outperforming existing approaches' in the video setting is unsubstantiated. Either include PSTNet in the video benchmark (if the source data are available, as the paper notes they are 'unavailable' — but then a claim of superiority should be explicitly qualified) or restrict the claim to the image-based baselines evaluated.
  3. [Abstract; Table 1] The abstract and conclusion claim that the method 'outperforms existing approaches', but Table 1 shows it is below several methods on ISTD+ (34.72 dB vs. 35.72 for HomoFormer and 35.46 for ShadowFormer) and below HomoFormer on SRD. The paper's own text ('Results on public and our benchmarks') says the method achieves 'competitive' performance on ISTD+/SRD. The claims should be aligned with the evidence: report the result as SOTA only on the proposed benchmark and for the video model, or acknowledge the public-benchmark trade-off explicitly.
minor comments (6)
  1. [Eq. (7)] The kernel definition is incomplete: 'k = 1/25 1_{5×5}' should specify that the 5×5 filter contains all ones and state explicitly that the channel-wise convolution is applied independently per channel.
  2. [Table 4] The 'Ours (frame-by-frame)' variant has a higher Ewarp (2.208) than OmniSR (2.032), while the video model achieves 1.775. Please add a sentence explaining whether the frame-by-frame variant is expected to be less temporally consistent and whether the difference is meaningful given the small benchmark size.
  3. [Table 1] The '–' entries (e.g., PhaSR on SRD, several methods on WildShadow-I) should be explicitly footnoted to indicate whether the result is unavailable or the method was not applicable, so that readers can distinguish 'not reported' from 'not evaluated'.
  4. [Our Video Dataset] The construction of WildShadow-I from 'INS, iCity, and Infinigen' is underspecified: it is unclear whether the images are sampled as single frames from the video clips or generated separately, and what the 'INS' split contains. Clarifying the sampling protocol and the exact training/test split would improve reproducibility.
  5. [Implementation Details] No code or dataset release URL is provided. Given that the paper introduces two new benchmarks (WildShadow and WildShadow-I), an availability statement is important for the community to reproduce and build upon the contribution.
  6. [Overall] The reference format is uneven (e.g., 'Smair 2024' is cited as (Smair 2024) but the reference list uses a different key formatting), and there are minor typographical issues in the extracted abstract. A careful proofread is recommended.

Circularity Check

0 steps flagged

No circular derivation: the method is independently ablated and checked on external benchmarks; same-author citations are design inheritance, not load-bearing reductions.

full rationale

The paper does not reduce any claimed result to a fitted parameter or to a self-citation chain. The core derivation is a LoRA fine-tuned Wan-Control video diffusion model with depth conditioning, a detail-injection module, and a shadow-mask-guided frequency-decomposed modulation. Each component is independently ablated in Table 3 (e.g., removing the detail injection module drops PSNR by 0.62 dB), so the contributions are not asserted by definition. The shadow mask M = 1 - X/(Y0_hat + epsilon) is computed from the input and the coarse prediction and is used only as guidance; the final output is a zero-initialized gated residual on the coarse decoded video, not a construction that forces the target. Public benchmarks ISTD+, SRD, and INS in Table 1 provide external grounding independent of the authors' own WildShadow datasets. The same-author citations (OmniSR for the rendering protocol, StableSR for the detail-injection inspiration) are non-load-bearing: those works also appear as baselines, and the design choices they inspire are quantitatively validated. The only substantive concern is external validity, not circularity: real-world evaluation (Fig. 5) is qualitative, and the paper itself states in its Limitations: 'our framework relies primarily on synthetic paired data for training, and how to effectively leverage large-scale real-world videos through self-supervised or weakly supervised learning remains an open problem.' The missing direct comparison with PSTNet also weakens the video-level 'outperforms existing approaches' claim, but that is a missing-experiment issue, not a circular step. Score 2 reflects minor self-citation without load-bearing circularity.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 0 invented entities

The contribution is empirical and introduces no new physical or conceptual entities. The main upstream commitments are the pretrained Wan-Control and DA3 models, hand-chosen hyperparameters (LoRA rank, FDM kernel), and the assumption that synthetic renderings transfer to real-world shadow removal.

free parameters (4)
  • LoRA rank / scaling factor = 64 / 64
    Chosen by hand; controlling adapter capacity. No rank sweep is reported, and the paper's performance could depend on this choice.
  • FDM low-pass kernel size = 5×5 mean filter
    Fixed channel-wise mean filter decomposes features into low/high frequency; the frequency split is central to the proposed modulation, but no ablation of kernel size is given.
  • Fixed text prompt
    The model is conditioned on the fixed prompt 'remove shadows from the video'; no prompt variations or null-text conditioning are explored.
  • Mask estimation epsilon = 1e-8
    Used in M = 1 - X/(Ŷ0 + ε) to avoid division by zero; minor numerical stability choice.
axioms (4)
  • domain assumption Pretrained Wan-Control video diffusion model provides useful spatiotemporal priors for shadow removal
    The whole approach freezes the VDM and only LoRA-adapts it; if the pretrained prior is unsuitable for shadow removal, the framework fails. Invoked in Section 'Latent Video Diffusion for Shadow Removal'.
  • domain assumption Depth Anything 3 depth estimates are reliable under shadowed, diverse lighting
    Eq. (2) uses DA3 on the shadowed input to supply geometry-aware conditioning; shadows can distort depth estimates, and inaccuracies would mislead the model.
  • ad hoc to paper Synthetic renderings (3D-FRONT, iCity, Infinigen) with only direct-illumination shadows removed are valid training targets for real-world shadow removal
    This is the core synthetic-to-real assumption in the 'Our Video Dataset' section. The real-world evaluation is qualitative only, so this assumption is not quantitatively validated.
  • domain assumption The coarse model output Ŷ0 is accurate enough that the mask M = 1 - X/(Ŷ0 + ε) localizes shadows
    The detail injection relies on this mask; if Ŷ0 contains artifacts or misalignments, FDM will modulate the wrong regions. Introduced in 'Shadow-mask-guided Detail Injection Module'.

pith-pipeline@v1.3.0-alltime-deepseek · 11692 in / 11533 out tokens · 100258 ms · 2026-08-01T00:29:10.798313+00:00 · methodology

0 comments
read the original abstract

Video shadow removal in the wild remains challenging due to complex illumination, diverse shadow appearances, and limited training data. Despite its importance to numerous vision and graphics applications, it remains largely unexplored in unconstrained real-world scenarios. To address this gap, we present WildShadowRemover, a framework that adapts a pretrained video diffusion model for robust video shadow removal via LoRA fine-tuning. To preserve fine image details while retaining the model's powerful generative prior, we augment the frozen VAE decoder with a detail injection module and introduce a shadow-mask-guided frequency-decomposed modulation module to selectively restore high-frequency textures while suppressing shadow artifacts. Monocular depth priors from Depth Anything 3 further provide geometry-aware guidance under challenging lighting conditions. We also construct WildShadow, a large-scale paired video shadow removal dataset and benchmark, covering diverse synthetic scenes. Extensive experiments demonstrate that our method outperforms existing approaches in shadow removal quality and temporal consistency, producing temporally coherent shadow-free videos with superior visual quality and strong generalization across challenging in-the-wild scenarios.

Figures

Figures reproduced from arXiv: 2607.26203 by Chi Wang, Cong Wang, Gang Xu, Jiamin Xu, Renshu Gu, Weiwei Xu, Zheng Dong.

Figure 1
Figure 1. Figure 1: Our proposed WildShadowRemover effectively removes complex shadows from in-the-wild videos. Given a shadowed input video (top), our method produces shadow￾free results (bottom) while preserving scene details and main￾taining temporal consistency across frames. Indoors (Raistrick et al. 2024), generate paired training data through physically based rendering, but primarily focus on indoor environments. As a … view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the proposed video shadow removal framework. A latent video diffusion model predicts a coarse shadow-free result, followed by a shadow-mask-guided detail injection module for detail restoration. Red arrows highlight enlarged regions showing the improved detail preservation of the final output. where T denotes the number of frames. Our framework per￾forms shadow removal in the latent space of a … view at source ↗
Figure 3
Figure 3. Figure 3: Our video dataset. Each row shows two represen￾tative frames from paired shadowed input and shadow-free ground-truth videos. rank and scaling factor are both set to 64. The per-GPU batch size is 1, resulting in a global batch size of 4. In the second stage, we freeze the first-stage DiT and Wan VAE backbone and optimize only the detail injection mod￾ules, which contain approximately 42.94M trainable parame… view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparison on INS, HSSD, and iCity. For this comparison, all methods are trained on our WildShadow-I training set. Our method achieves more effective shadow removal while preserving color consistency and fine-grained scene details. Red arrows indicate regions with clearer shadow removal. method with OmniSR, which achieves the second-best per￾formance on our WildShadow-I benchmark. When trained … view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative comparison on real-world DL3DV [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗

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

124 extracted references · 16 linked inside Pith

  1. [1]

    2018 , isbn =

    Akenine-Mller, Tomas and Haines, Eric and Hoffman, Naty , title =. 2018 , isbn =

    2018

  2. [2]

    2018 , url =

    Blender - a 3D modelling and rendering package , author =. 2018 , url =

    2018

  3. [3]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Leave-one-out kernel optimization for shadow detection and removal , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2017 , publisher=

    2017

  4. [4]

    IEEE transactions on pattern analysis and machine intelligence , volume=

    Paired regions for shadow detection and removal , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2012 , publisher=

    2012

  5. [5]

    2010 IEEE Computer Society conference on computer vision and pattern recognition , pages=

    Learning to recognize shadows in monochromatic natural images , author=. 2010 IEEE Computer Society conference on computer vision and pattern recognition , pages=. 2010 , organization=

    2010

  6. [6]

    arXiv preprint arXiv:2312.16256 , volume=

    Dl3dv-10k: A large-scale scene dataset for deep learning-based 3d vision , author=. arXiv preprint arXiv:2312.16256 , volume=

    Pith/arXiv arXiv

  7. [7]

    International Conference on Learning Representations (ICLR) , year=

    LoRA: Low-Rank Adaptation of Large Language Models , author=. International Conference on Learning Representations (ICLR) , year=

  8. [9]

    arXiv preprint arXiv:2605.06658 , year=

    Relit-LiVE: Relight Video by Jointly Learning Environment Video , author=. arXiv preprint arXiv:2605.06658 , year=

    Pith/arXiv arXiv

  9. [10]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Objectmover: Generative object movement with video prior , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  10. [11]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Detail-preserving latent diffusion for stable shadow removal , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  11. [12]

    Advances in neural information processing systems , volume=

    Video diffusion models , author=. Advances in neural information processing systems , volume=

  12. [13]

    arXiv preprint arXiv:2503.20314 , year=

    Wan: Open and advanced large-scale video generative models , author=. arXiv preprint arXiv:2503.20314 , year=

    Pith/arXiv arXiv

  13. [14]

    arXiv preprint arXiv:2511.10647 , year=

    Depth anything 3: Recovering the visual space from any views , author=. arXiv preprint arXiv:2511.10647 , year=

    Pith/arXiv arXiv

  14. [15]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Infinite photorealistic worlds using procedural generation , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  15. [16]

    , author=

    3d gaussian splatting for real-time radiance field rendering. , author=. ACM Trans. Graph. , volume=

  16. [17]

    2024 , howpublished =

    iCity: Procedural City Generator for Blender , author =. 2024 , howpublished =

    2024

  17. [18]

    2306.11290 , archivePrefix=

    arXiv preprint , year=. 2306.11290 , archivePrefix=

    Pith/arXiv arXiv

  18. [19]

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

    No Shadow Left Behind: Removing Objects and Their Shadows Using Approximate Lighting and Geometry , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year=

  19. [20]

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

    Raistrick, Alexander and Mei, Lingjie and Kayan, Karhan and Yan, David and Zuo, Yiming and Han, Beining and Wen, Hongyu and Parakh, Meenal and Alexandropoulos, Stamatis and Lipson, Lahav and Ma, Zeyu and Deng, Jia , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2024 , pages =

    2024

  20. [21]

    Computer Vision--ECCV 2010: 11th European Conference on Computer Vision, Heraklion, Crete, Greece, September 5-11, 2010, Proceedings, Part II 11 , pages=

    Detecting ground shadows in outdoor consumer photographs , author=. Computer Vision--ECCV 2010: 11th European Conference on Computer Vision, Heraklion, Crete, Greece, September 5-11, 2010, Proceedings, Part II 11 , pages=. 2010 , organization=

    2010

  21. [22]

    2011 international conference on computer vision , pages=

    What characterizes a shadow boundary under the sun and sky? , author=. 2011 international conference on computer vision , pages=. 2011 , organization=

    2011

  22. [23]

    2014 IEEE conference on computer vision and pattern recognition , pages=

    Automatic feature learning for robust shadow detection , author=. 2014 IEEE conference on computer vision and pattern recognition , pages=. 2014 , organization=

    2014

  23. [24]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

    Shadow optimization from structured deep edge detection , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

  24. [25]

    Computer Vision--ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part VI 14 , pages=

    Large-scale training of shadow detectors with noisily-annotated shadow examples , author=. Computer Vision--ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part VI 14 , pages=. 2016 , organization=

    2016

  25. [26]

    Proceedings of the IEEE International Conference on Computer Vision , pages=

    Shadow detection with conditional generative adversarial networks , author=. Proceedings of the IEEE International Conference on Computer Vision , pages=

  26. [27]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Mask-shadowgan: Learning to remove shadows from unpaired data , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  27. [28]

    Proceedings of the European Conference on Computer Vision (ECCV) , pages=

    Bidirectional feature pyramid network with recurrent attention residual modules for shadow detection , author=. Proceedings of the European Conference on Computer Vision (ECCV) , pages=

  28. [29]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Stacked conditional generative adversarial networks for jointly learning shadow detection and shadow removal , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  29. [30]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Argan: Attentive recurrent generative adversarial network for shadow detection and removal , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  30. [31]

    Proceedings of the European Conference on Computer Vision (ECCV) , pages=

    A+ D Net: Training a shadow detector with adversarial shadow attenuation , author=. Proceedings of the European Conference on Computer Vision (ECCV) , pages=

  31. [32]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Distraction-aware shadow detection , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  32. [33]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Instance shadow detection , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  33. [34]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Mitigating intensity bias in shadow detection via feature decomposition and reweighting , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  34. [35]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Single-stage instance shadow detection with bidirectional relation learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  35. [36]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=

    Instance shadow detection with a single-stage detector , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , volume=. 2022 , publisher=

    2022

  36. [37]

    2022 IEEE International Conference on Multimedia and Expo (ICME) , pages=

    RMLANet: random multi-level attention network for shadow detection , author=. 2022 IEEE International Conference on Multimedia and Expo (ICME) , pages=. 2022 , organization=

    2022

  37. [38]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    SILT: Shadow-aware Iterative Label Tuning for Learning to Detect Shadows from Noisy Labels , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  38. [39]

    Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition , pages=

    A multi-task mean teacher for semi-supervised shadow detection , author=. Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition , pages=

  39. [40]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Bijective mapping network for shadow removal , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  40. [41]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Canet: A context-aware network for shadow removal , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  41. [42]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Auto-exposure fusion for single-image shadow removal , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  42. [43]

    IEEE Transactions on Neural Networks and Learning Systems , year=

    A Shadow Imaging Bilinear Model and Three-Branch Residual Network for Shadow Removal , author=. IEEE Transactions on Neural Networks and Learning Systems , year=

  43. [44]

    arXiv preprint arXiv:2302.01650 , year=

    Shadowformer: Global context helps image shadow removal , author=. arXiv preprint arXiv:2302.01650 , year=

    Pith/arXiv arXiv

  44. [45]

    IEEE Transactions on Multimedia , year=

    A Decoupled Multi-Task Network for Shadow Removal , author=. IEEE Transactions on Multimedia , year=

  45. [46]

    IEEE transactions on pattern analysis and machine intelligence , volume=

    Direction-aware spatial context features for shadow detection and removal , author=. IEEE transactions on pattern analysis and machine intelligence , volume=. 2019 , publisher=

    2019

  46. [47]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Towards ghost-free shadow removal via dual hierarchical aggregation network and shadow matting gan , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  47. [48]

    Proceedings of the AAAI conference on artificial intelligence , volume=

    Efficient model-driven network for shadow removal , author=. Proceedings of the AAAI conference on artificial intelligence , volume=

  48. [49]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

    Deshadownet: A multi-context embedding deep network for shadow removal , author=. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages=

  49. [50]

    Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XI 16 , pages=

    From shadow segmentation to shadow removal , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XI 16 , pages=. 2020 , organization=

    2020

  50. [51]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    From shadow generation to shadow removal , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  51. [52]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Shadowdiffusion: When degradation prior meets diffusion model for shadow removal , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  52. [53]

    CoRR , year=

    Adam: A Method for Stochastic Optimization , author=. CoRR , year=

  53. [54]

    Automatic differentiation in pytorch , author=. , year=

  54. [55]

    arXiv preprint arXiv:1608.03983 , year=

    Sgdr: Stochastic gradient descent with warm restarts , author=. arXiv preprint arXiv:1608.03983 , year=

    Pith/arXiv arXiv

  55. [56]

    Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XXV 16 , pages=

    Learning enriched features for real image restoration and enhancement , author=. Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XXV 16 , pages=. 2020 , organization=

    2020

  56. [57]

    IEEE TIP , volume=

    Image quality assessment: from error visibility to structural similarity , author=. IEEE TIP , volume=

  57. [58]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Shadowneus: Neural sdf reconstruction by shadow ray supervision , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  58. [59]

    Proceedings of the IEEE international conference on computer vision , pages=

    Mask r-cnn , author=. Proceedings of the IEEE international conference on computer vision , pages=

  59. [60]

    2010 20th International Conference on Pattern Recognition , pages=

    Improved shadow removal for robust person tracking in surveillance scenarios , author=. 2010 20th International Conference on Pattern Recognition , pages=. 2010 , organization=

    2010

  60. [61]

    Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

    Learning intrinsic image decomposition from watching the world , author=. Proceedings of the IEEE conference on computer vision and pattern recognition , pages=

  61. [62]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Learning physics-guided face relighting under directional light , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  62. [63]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Intrinsicnerf: Learning intrinsic neural radiance fields for editable novel view synthesis , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  63. [64]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Shadow removal via shadow image decomposition , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  64. [65]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Adaptive Illumination Mapping for Shadow Detection in Raw Images , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  65. [66]

    International Journal of Computer Vision , pages=

    Learning to Remove Shadows from a Single Image , author=. International Journal of Computer Vision , pages=. 2023 , publisher=

    2023

  66. [67]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Dc-shadownet: Single-image hard and soft shadow removal using unsupervised domain-classifier guided network , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  67. [68]

    arXiv preprint arXiv:2302.05361 , year=

    Leveraging Inpainting for Single-Image Shadow Removal , author=. arXiv preprint arXiv:2302.05361 , year=

    Pith/arXiv arXiv

  68. [69]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    3d-front: 3d furnished rooms with layouts and semantics , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  69. [70]

    International Journal of Computer Vision , pages=

    3d-future: 3d furniture shape with texture , author=. International Journal of Computer Vision , pages=. 2021 , publisher=

    2021

  70. [71]

    Advances in Neural Information Processing Systems , volume=

    Atiss: Autoregressive transformers for indoor scene synthesis , author=. Advances in Neural Information Processing Systems , volume=

  71. [72]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Vision transformers for dense prediction , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  72. [73]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Omnidata: A scalable pipeline for making multi-task mid-level vision datasets from 3d scans , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  73. [74]

    arXiv preprint arXiv:2304.07193 , year=

    Dinov2: Learning robust visual features without supervision , author=. arXiv preprint arXiv:2304.07193 , year=

    Pith/arXiv arXiv

  74. [75]

    ACM Transactions on Graphics (TOG) , volume=

    Materialistic: Selecting Similar Materials in Images , author=. ACM Transactions on Graphics (TOG) , volume=. 2023 , publisher=

    2023

  75. [76]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Swin transformer: Hierarchical vision transformer using shifted windows , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

  76. [77]

    Medical Image Computing and Computer-Assisted Intervention--MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18 , pages=

    U-net: Convolutional networks for biomedical image segmentation , author=. Medical Image Computing and Computer-Assisted Intervention--MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18 , pages=. 2015 , organization=

    2015

  77. [78]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Uformer: A general u-shaped transformer for image restoration , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  78. [79]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Inverse rendering for complex indoor scenes: Shape, spatially-varying lighting and svbrdf from a single image , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  79. [80]

    European Conference on Computer Vision , pages=

    Physically-based editing of indoor scene lighting from a single image , author=. European Conference on Computer Vision , pages=. 2022 , organization=

    2022

  80. [81]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Adding conditional control to text-to-image diffusion models , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

Showing first 80 references.