Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Survey on Single-Image Reflection Removal using Deep Learning Techniques

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This survey argues that deep learning research on single-image reflection removal can be organized by image-formation hypothesis and network stage count, and that the field's progress is currently bottlenecked by scarce, low-quality data…

desk verdict A useful but uneven survey: the stage-based taxonomy and dataset summary are genuinely handy, yet the 28-paper selection is not auditable and the stagnation claim outruns the evidence. read the letter →

arxiv 2502.08836 v1 pith:NPKXFYFA submitted 2025-02-12 cs.CV

Kangning Yang , Huiming Sun , Jie Cai , Lan Fu , Jiaming Ding , Jinlong Li , Chiu Man Ho , Zibo Meng This is my paper · ORCID
classification cs.CV
keywords reflectionremovalsingle-imageseparationdeeplearningsurveyimagelayerdecompositionsyntheticandrealdatasetsevaluationmetricscomputervision
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 survey tries to establish that deep learning research on single-image reflection removal, published in eight major venues from 2017 to 2025, can be comprehensively organized by two axes: the mathematical hypothesis of how a reflected image is formed, and the number of stages in the network architecture. It argues that the field's progress is currently held back less by network design than by the scarcity of large, diverse, real-world training and test datasets, and by an unclear task definition that mixes reflection removal with background reconstruction. A sympathetic reader would care because the survey provides a structured map of 28 methods, the datasets they use, and the metrics used to judge them, making it easier to position future work and to see where the bottlenecks are.

What carries the argument

The organizing device is a two-axis framework. The first axis is the formation hypothesis: how a mixed image $I$ is modeled from a transmission layer $T$ and a reflection layer $R$, ranging from linear $I = T + R$ and $I = \alpha T + \beta R$ to non-linear $I = W \circ T + R$ and $I = T + R + \Phi(T,R)$. The second axis is the architectural stage count: single-stage networks that map $I$ to $T$ and/or $R$ directly, two-stage networks that first estimate an intermediate (edge map, reflection, or absorption coefficient) and then reconstruct the transmission, and multi-stage recurrent cascades that refine estimates iteratively. This framework is what allows the survey to compare heterogeneous methods and to attribute the field's slowdown to data rather than to model capacity.

What would settle it

A concrete check: run the same search query across the same eight venues for 2017-2025 and compare the resulting set to the 28 papers the survey says it includes. If any additional relevant paper appears, or if the survey's own references list papers from those venues that are not among the 28, the comprehensiveness claim is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that every deep learning approach to single-image reflection removal fits into a taxonomy of single-stage, two-stage, and multi-stage architectures, and that the choice of image-formation hypothesis—from the simple linear superposition $I = T + R$, to blending-scalar variants $I = \alpha T + \beta R$, to non-linear $\alpha$-matte models $I = W \circ T + R$ and residual formulations $I = T + R + \Phi(T,R)$—determines how the network is structured and what it can separate. The paper further claims that the current state of the field is characterized by stagnation in network evolution because small datasets make simple models like UNet competitive with elaborate architectures, and that the remedy is larger real-world datasets, clearer task definitions, and the eventual integration of multimodal and foundation-model guidance. This is asserted as a synthesis of the surveyed literature rather than as a new algorithm.

Load-bearing premise

The survey's comprehensiveness rests on the assumption that its keyword search of eight venues, followed by manual screening, captured a representative sample of the field, but the paper does not list the 28 selected papers, so a reader cannot verify that no relevant work was missed.

Editorial extensions

If this is right

  • If the taxonomy is accepted, new reflection-removal papers can be described by their formation hypothesis and span count, making method comparison more systematic.
  • Because the survey argues that small datasets let simple UNet-style models reach state-of-the-art, it follows that publishing larger paired real-world datasets would do more for progress than further architectural innovation.
  • The paper's call for a unified evaluation framework implies that cross-dataset benchmarks on SIR2, CDR, and RRW would settle which methods actually generalize.
  • The survey's identification of task ambiguity suggests that deciding whether severe reflections require inpainting, not just layer separation, is a prerequisite for meaningful metric comparison.

Reading between the lines

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

  • The paper does not compare benchmark numbers, so a natural next step would be to run the 28 surveyed methods on a shared test set; this survey provides the roster but not the rankings.
  • Reading the survey's stagnation argument as a hypothesis, one testable prediction is that a recent model trained on RRW and evaluated on SIR2 will outperform the same model trained on CEIL by a larger margin than the total gains from any architectural change since 2017.
  • The survey's emphasis on foundation models suggests that language-guided separation, as in the 2024 paper it covers, may become the dominant paradigm; that is an extrapolation the authors hint at but do not develop.
  • Because the search was limited to English-language venues, the survey's coverage claim implicitly assumes that no significant SIRR work appears in other venues or languages; a reader planning to rely on its completeness should verify against a broader search.
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. This manuscript is a survey of deep-learning-based single-image reflection removal (SIRR). It describes a bibliographic search restricted to eight high-impact venues, organizes the selected literature into single-stage, two-stage, and multi-stage approaches, presents mathematical formation models (linear, blending-scalar, and non-linear), summarizes public datasets and evaluation metrics, and ends with challenges, future directions, and limitations. The authors claim a comprehensive, critically assessed review and identify a three-fold contribution: summarizing recent work, outlining hypotheses/techniques/datasets/metrics, and highlighting challenges and opportunities.

Significance. If the survey were fully reproducible and internally consistent, it would be a useful entry point for researchers entering SIRR. The paper's organization by architectural stage, its compilation of formation hypotheses, and its dataset table are convenient and generally accurate. The explicit discussion of limitations and future directions is also constructive. However, the main value of a survey in this setting is its comprehensiveness and reliability, and both are currently weakened by the unverifiable paper-selection procedure and by specific factual inconsistencies in the dataset section. These issues do not require new experiments to fix, but they do require substantive revision, so the appropriate outcome is major revision rather than rejection.

major comments (4)
  1. [Section 2; Table 1] The paper-selection procedure is not auditable as reported. The authors state that 28 papers remained after their search, but they never list those 28 papers, and Table 1 tabulates only 19 methods, so a reader cannot verify which papers were included or why. Section 6.1 then concludes that "academic research in this field is stagnating" based on this unverifiable sample; Section 6.3 also admits that relevant work may have been missed due to the chosen keywords and databases. Because the abstract's "comprehensive review" claim and the stagnation conclusion depend on the sample, this is a load-bearing weakness. The authors should provide a complete list of the 28 retained papers (e.g., as an appendix), state the databases searched and the inclusion/exclusion criteria, and either justify the venue restriction or reframe the review as covering a representative sample rather than a comprehensive corpus.
  2. [Section 5.2; Table 2] The text and Table 2 directly contradict each other on the composition of the SIR2 and CEIL datasets. The text says that "SIR2 [10] and CEIL [13], are larger and include both synthetic and real-world images," while Table 2 labels SIR2 as "Real" only and CEIL as "Syn" only. This is a factual inconsistency in one of the paper's central reference tables and must be corrected, with the true composition of each dataset verified against the original sources.
  3. [Table 2; Reference [34]] The year for the CDR dataset is inconsistent: Table 2 lists CDR as 2021, but reference [34] is a CVPR 2022 paper. The authors should verify the year and the citation for CDR, and they should audit the other table entries (e.g., the SIR2+ entry, which cites reference [35] that appears to duplicate reference [11]) to ensure that every dataset's year and source match the cited publication.
  4. [Abstract; Section 4; Section 6.1] The abstract promises a "critical assessment" of single-stage and two-stage methods, but Section 4 is largely a descriptive summary of each method, with no explicit comparative strengths/weaknesses analysis. Similarly, the claim in Section 6.1 that the field is "stagnating" is stronger than the restricted sample supports, especially since Section 6.3 concedes that relevant work may have been missed. The manuscript should either add a comparative critical discussion (e.g., a strengths/limitations table or a structured comparison of methods) or soften the claims to match the descriptive level actually provided.
minor comments (5)
  1. [Section 1] There is a typo in the Introduction: "with the aim to of presenting" should read "with the aim of presenting."
  2. [Section 4.3] The text refers to "DBN" in the first sentence of Section 4.3, but Table 1 and reference [20] use "BDN"; please use the abbreviation consistently.
  3. [Section 5.2] The sentence describing CDR is grammatically broken: "The CDR dataset is categorized according to reflection types and contains images with perfect alignment between the mixed and transmission images includes misaligned raw flash/ambient images." This should be rewritten for clarity.
  4. [Section 5.3.2] The contextual loss formula in Section 5.3 is presented without an equation number and with an ambiguous "max_j CX_ij" expression; please add a numbering and define the index notation.
  5. [References] References [11] and [35] appear to cite the same TPAMI paper; please merge or disambiguate them to avoid duplication.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: this is a descriptive survey with no derivation chain, fitted parameters, or self-citation-driven claims.

full rationale

This paper is a literature survey, not a derivation or prediction pipeline. Its contributions are descriptive: it summarizes single-image reflection removal methods, organizes them by architecture stage, lists datasets and loss functions, and discusses challenges. There is no equation whose output is defined in terms of its own conclusion, no fitted parameter that is later renamed as a prediction, and no load-bearing self-citation chain. The paper's central claims are that it provides a comprehensive review and that academic research in this field is stagnating; both are interpretive summaries of the 28 papers the authors selected. The selection methodology is an input to the survey, not a quantity derived from its own output, so even if the sample is incomplete or unverifiable, that is a completeness and reproducibility limitation, not circularity. The paper explicitly acknowledges in Section 6.3 that some related research may have been missed, which further confirms that the authors are not presenting the selection as a logically forced result. None of the enumerated circularity patterns apply: the paper does not define its taxonomy in terms of its conclusions, does not import a uniqueness theorem from its own authors, and does not smuggle an ansatz in via citation. The survey is self-contained as a descriptive document, and its value or correctness depends on coverage and accuracy of reporting, not on circular reasoning.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The survey does not fit parameters or introduce new entities. It relies on the representativeness of its literature sample, the validity of its taxonomy, and accurate historical attributions; none of these are given independent proof in the paper.

assumptions (3)
  • ad hoc to paper The bibliographic search over 8 venues and the given keyword query, followed by manual screening, yields a representative sample of deep learning-based SIRR research.
    This is the survey's core methodological premise, stated in Section 2; if the sample is biased or incomplete, the survey's comprehensiveness claim fails.
  • domain assumption The stage-based taxonomy (single-stage, two-stage, multi-stage) is a valid organizing framework for SIRR methods.
    The survey categorizes all methods into these three groups in Section 4; there is no formal argument that these categories are exhaustive or mutually exclusive.
  • domain assumption The historical attribution that Fan et al. [13] introduced the first neural network model for SIRR in 2017 is correct.
    Stated in Section 2 without detailed evidence; if earlier neural network work exists, the survey's timeline is wrong.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Survey on Single-Image Reflection Removal using Deep Learning Techniques." pith.science (2026). https://pith.science/paper/NPKXFYFA

@misc{pith2026250208836,
  author       = {Pith},
  title        = {Pith review of: Survey on Single-Image Reflection Removal using Deep Learning Techniques},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NPKXFYFA}},
  note         = {Machine review of arXiv:2502.08836}
}
read the original abstract

The phenomenon of reflection is quite common in digital images, posing significant challenges for various applications such as computer vision, photography, and image processing. Traditional methods for reflection removal often struggle to achieve clean results while maintaining high fidelity and robustness, particularly in real-world scenarios. Over the past few decades, numerous deep learning-based approaches for reflection removal have emerged, yielding impressive results. In this survey, we conduct a comprehensive review of the current literature by focusing on key venues such as ICCV, ECCV, CVPR, NeurIPS, etc., as these conferences and journals have been central to advances in the field. Our review follows a structured paper selection process, and we critically assess both single-stage and two-stage deep learning methods for reflection removal. The contribution of this survey is three-fold: first, we provide a comprehensive summary of the most recent work on single-image reflection removal; second, we outline task hypotheses, current deep learning techniques, publicly available datasets, and relevant evaluation metrics; and third, we identify key challenges and opportunities in deep learning-based reflection removal, highlighting the potential of this rapidly evolving research area.

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

  1. OpenRR-1k: A Scalable Dataset for Real-World Reflection Removal

    cs.CV 2025-06 conditional novelty 6.0 of 10

    OpenRR-1k is a 1,000-pair real-world reflection removal dataset generated by off-the-shelf AI removal plus manual refinement, with benchmarks showing fine-tuning gains.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages · cited by 1 Pith paper

  1. [34]

    Single image reflection re moval with physically-based training images,

    S. Kim, Y . Huo, and S. E. Y oon, “Single image reflection re moval with physically-based training images,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 5164–5173

    work page 2020

  2. [10]

    User assisted separation of reflec tions from a single image using a sparsity prior,

    A. Levin and Y . Weiss, “User assisted separation of reflec tions from a single image using a sparsity prior,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 29, no. 9, pp. 1647–1654, 2007

    work page 2007

  3. [13]

    Depth of field guid ed reflection removal,

    R. Wan, B. Shi, T. A. Hwee, and A. C. Kot, “Depth of field guid ed reflection removal,” in 2016 IEEE International Conference on Image Processing. IEEE, 2016, pp. 21–25

    work page 2016

  4. [35]

    Revisiting single image reflection removal in the wil d,

    Y . Zhu, X. Fu, P . T. Jiang, H. Zhang, Q. Sun, J. Chen, Z. J. Z ha, and B. Li, “Revisiting single image reflection removal in the wil d,” in Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 25468–25478

    work page 2024

  5. [11]

    Single image layer separation usin g relative smoothness,

    Y . Li and M. S. Brown, “Single image layer separation usin g relative smoothness,” in Proceedings of the IEEE conference on computer vi- sion and pattern recognition, 2014, pp. 2752–2759

    work page 2014

  6. [1]

    Over the years, various techniques have been proposed to solve the SIRR problem

    INTRODUCTION Single-image reflection removal (SIRR) is a critical task in image processing, focusing on recovering the true scene behind reflections from reflective surfaces (e.g., transpar ent glasses). Over the years, various techniques have been proposed to solve the SIRR problem. Traditional methods typically re- lied on a non-learning paradigm. Since th...

    work page 2015

  7. [2]

    single image reflection removal

    METHODOLOGY To conduct a comprehensive and focused survey of the relevant literature, we strategically concentrated our bi bli- ographic search on key conferences and journals that are widely recognized for publishing state-of-the-art high-i mpact research in computer vision and artificial intelligence. In stead of conducting a broad search across all avai...

    work page 2017

  8. [3]

    To this end, researchers have proposed various hy- potheses

    MA THEMA TICAL HYPOTHESIS As previously mentioned, SIRR is inherently an ill-posed problem. To this end, researchers have proposed various hy- potheses. 3.1. Linear Hypothesis The linear hypothesis posits that a captured image I is perceived as the superimposition of a transmission layer an d a reflection layer, a concept inspired by the human visual sys -...

    work page

Show all 43 references
  1. [4]

    The learning objective is introduced i n Section 4.4, with a comparative analysis of models in Table 1

    REFLECTION REMOV AL APPROACHES We first present one-stage SIRR approaches in Sec- tion 4.1, followed by two-stage and multi-stage approaches in Sections 4.2 and 4.3. The learning objective is introduced i n Section 4.4, with a comparative analysis of models in Table 1. 4.1. Sin...

    2018

  2. [5]

    Data Acquisition The datasets for SIRR are a critical aspect of developing effective deep learning models

    DA TASETS & EV ALUA TION METRICS 5.1. Data Acquisition The datasets for SIRR are a critical aspect of developing effective deep learning models. These datasets vary in size , image source, and the type of annotations provided. In gen- eral, they can be categorized into two mai...

    2017

  3. [6]

    DISCUSSION 6.1. Challenges in current SIRR research One of the biggest challenges in SIRR research is the lack of large, high-quality training datasets that represent a v ari- ety of reflection types across different surfaces and lighti ng conditions. Reflection removal relies o...

  4. [7]

    We first describe the methodology used for our bibliographic search

    CONCLUSION This article presents an overview of SIRR utilizing deep learning methods. We first describe the methodology used for our bibliographic search. Then, we provide a comprehensive analysis of research categorized by single-stage, two-sta ge, and multi-stage approaches. ...

  5. [8]

    Separating reflections from a single image using local features,

    A. Levin, A. Zomet, and Y . Weiss, “Separating reflections from a single image using local features,” in Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004. IEEE, 2004, vol. 1, pp. I–I

    2004

  6. [9]

    Learning to perceive tr ansparency from the statistics of natural scenes,

    A. Levin, A. Zomet, and Y . Weiss, “Learning to perceive tr ansparency from the statistics of natural scenes,” Advances in Neural Information Processing Systems, vol. 15, 2002

    2002

  7. [12]

    Fast single i m- age reflection suppression via convex optimization,

    Y . Y ang, W. Ma, Y . Zheng, J. F. Cai, and W. Xu, “Fast single i m- age reflection suppression via convex optimization,” in Proceedings of the IEEE/CVF conference on computer vision and pattern reco gnition, 2019, pp. 8141–8149

    2019

  8. [14]

    Reflec tion re- moval using ghosting cues,

    Y . Shih, D. Krishnan, F. Durand, and W. T. Freeman, “Reflec tion re- moval using ghosting cues,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 3193–3201

    2015

  9. [15]

    Robust single image reflection removal against adversaria l attacks,

    Z. Song, Z. Zhang, K. Zhang, W. Luo, Z. Fan, W. Ren, and J. Lu , “Robust single image reflection removal against adversaria l attacks,” in Proceedings of the IEEE/CVF Conference on Computer Vision a nd Pattern Recognition, 2023, pp. 24688–24698

    2023

  10. [16]

    Single image reflec- tion removal exploiting misaligned training data and netwo rk enhance- ments,

    K. Wei, J. Y ang, Y . Fu, D. Wipf, and H. Huang, “Single image reflec- tion removal exploiting misaligned training data and netwo rk enhance- ments,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 8178–8187

    2019

  11. [17]

    Benchm ark- ing single-image reflection removal algorithms,

    R. Wan, B. Shi, L. Y . Duan, A. H. Tan, and A. C. Kot, “Benchm ark- ing single-image reflection removal algorithms,” in Proceedings of the IEEE International Conference on Computer Vision , 2017, pp. 3922– 3930

    2017

  12. [18]

    Be nch- marking single-image reflection removal algorithms,

    R. Wan, B. Shi, H. Li, Y . Hong, L. Y . Duan, and A. C. Kot, “Be nch- marking single-image reflection removal algorithms,” IEEE Transac- tions on Pattern Analysis and Machine Intelligence , vol. 45, no. 2, pp. 1424–1441, 2022

    2022

  13. [19]

    Single-image reflection removal using deep lear ning: a systematic review,

    A. Amanlou, A. A. Suratgar, J. Tavoosi, A. Mohammadzade h, and A. Mosavi, “Single-image reflection removal using deep lear ning: a systematic review,” IEEE Access, vol. 10, pp. 29937–29953, 2022

    2022

  14. [20]

    A generic de ep archi- tecture for single image reflection removal and image smooth ing,

    Q. Fan, J. Y ang, G. Hua, B. Chen, and D. Wipf, “A generic de ep archi- tecture for single image reflection removal and image smooth ing,” in Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 3238–3247

    2017

  15. [21]

    Single image reflection sep aration with perceptual losses,

    X. Zhang, R. Ng, and Q. Chen, “Single image reflection sep aration with perceptual losses,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 4786–4794

    2018

  16. [22]

    Trash or treasure? an interactive dual -stream strategy for single image reflection separation,

    Q. Hu and X. Guo, “Trash or treasure? an interactive dual -stream strategy for single image reflection separation,” Advances in Neural Information Processing Systems, vol. 34, pp. 24683–24694, 2021

    2021

  17. [23]

    Single image reflection separation via component synergy,

    Q. Hu and X. Guo, “Single image reflection separation via component synergy,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 13138–13147

    2023

  18. [24]

    Reflection scene separation from a single image,

    R. Wan, B. Shi, H. Li, L. Y . Duan, and A. C. Kot, “Reflection scene separation from a single image,” in Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition , 2020, pp. 2398– 2406

    2020

  19. [25]

    Two-stage s ingle image reflection removal with reflection-aware guidance,

    Y . Li, M. Liu, Y . Yi, Q. Li, D. Ren, and W. Zuo, “Two-stage s ingle image reflection removal with reflection-aware guidance,” Applied In- telligence, vol. 53, no. 16, pp. 19433–19448, 2023

    2023

  20. [26]

    Crrn: M ulti- scale guided concurrent reflection removal network,

    R. Wan, B. Shi, L. Y . Duan, A. H. Tan, and A. C. Kot, “Crrn: M ulti- scale guided concurrent reflection removal network,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recogn ition, 2018, pp. 4777–4785

    2018

  21. [27]

    Seeing deeply and bi direction- ally: A deep learning approach for single image reflection re moval,

    J. Y ang, D. Gong, L. Liu, and Q. Shi, “Seeing deeply and bi direction- ally: A deep learning approach for single image reflection re moval,” in Proceedings of the european conference on computer vision , 2018, pp. 654–669

    2018

  22. [28]

    Single i mage reflection removal through cascaded refinement,

    C. Li, Y . Y ang, K. He, S. Lin, and J. E. Hopcroft, “Single i mage reflection removal through cascaded refinement,” in Proceedings of the IEEE/CVF conference on computer vision and pattern reco gnition, 2020, pp. 3565–3574

    2020

  23. [29]

    Single im age reflection removal beyond linearity,

    Q. Wen, Y . Tan, J. Qin, W. Liu, G. Han, and S. He, “Single im age reflection removal beyond linearity,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2019, pp. 3771–3779

    2019

  24. [30]

    Locat ion- aware single image reflection removal,

    Z. Dong, K. Xu, Y . Y ang, H. Bao, W. Xu, and R. W. Lau, “Locat ion- aware single image reflection removal,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 5017–5026

    2021

  25. [31]

    Single image reflection removal with absorption effect,

    Q. Zheng, B. Shi, J. Chen, X. Jiang, L. Y . Duan, and A. C. Ko t, “Single image reflection removal with absorption effect,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni tion, 2021, pp. 13395–13404

    2021

  26. [32]

    Co rrn: Cooperative reflection removal network,

    R. Wan, B. Shi, H. Li, L. Y Duan, A. H. Tan, and A. C. Kot, “Co rrn: Cooperative reflection removal network,” IEEE transactions on pat- tern analysis and machine intelligence, vol. 42, no. 12, pp. 2969–2982, 2019

    2019

  27. [33]

    Deep -masking generative network: A unified framework for background rest oration from superimposed images,

    X. Feng, W. Pei, Z. Jia, F. Chen, D. Zhang, and G. Lu, “Deep -masking generative network: A unified framework for background rest oration from superimposed images,” IEEE Transactions on Image Processing, vol. 30, pp. 4867–4882, 2021

    2021

  28. [36]

    Languag e-guided image reflection separation,

    H. Zhong, Y . Hong, S. Weng, J. Liang, and B. Shi, “Languag e-guided image reflection separation,” in Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition , 2024, pp. 24913– 24922

    2024

  29. [37]

    Single i mage re- flection removal with edge guidance, reflection classifier, a nd recurrent decomposition,

    Y . C. Chang, C. N. Lu, C. C. Cheng, and W. C. Chiu, “Single i mage re- flection removal with edge guidance, reflection classifier, a nd recurrent decomposition,” in Proceedings of the IEEE/CVF winter conference on applications of computer vision , 2021, pp. 2033–2042

    2021

  30. [38]

    V -desirr: V ery fast deep embedded single image reflection removal,

    B. Prasad, L. R. Boregowda, K. Mitra, S. Chowdhury, et al ., “V -desirr: V ery fast deep embedded single image reflection removal,” in Proceed- ings of the IEEE/CVF International Conference on Computer V ision, 2021, pp. 2390–2399

    2021

  31. [39]

    Perceptual losse s for real-time style transfer and super-resolution,

    J. Johnson, A. Alahi, and L. Fei-Fei, “Perceptual losse s for real-time style transfer and super-resolution,” in Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, Oct ober 11- 14, 2016, Proceedings, Part II 14 . Springer, 2016, pp. 694–711

    2016

  32. [40]

    Generative adversar ial nets,

    I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. War de-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversar ial nets,” Advances in neural information processing systems , vol. 27, 2014

    2014

  33. [41]

    A categorized reflection removal dataset with diverse real-w orld scenes,

    C. Lei, X. Huang, C. Qi, Y . Zhao, W. Sun, Q. Y an, and Q. Chen , “A categorized reflection removal dataset with diverse real-w orld scenes,” in Proceedings of the IEEE/CVF Conference on Computer Vision a nd Pattern Recognition, 2022, pp. 3040–3048

    2022

  34. [42]

    Benchmarking single-image reflection removal algorithms ,

    R. Wan, B. Shi, H. Li, Y . Hong, L. Y . Duan, and A. C. Kot, “Benchmarking single-image reflection removal algorithms ,” in IEEE Transacations on Pattern Analysis and Machine Intelligence, 2022

    2022

  35. [43]

    Backg round scene recovery from an image looking through colored glass,

    C. Wang, D. Xu, R. Wan, B. He, B. Shi, and L. Y . Duan, “Backg round scene recovery from an image looking through colored glass, ” IEEE Transactions on Multimedia, 2022

    2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.