Pith. sign in

REVIEW 4 major objections 5 minor 68 references

Towards Color-Faithful Low-Light Image Enhancement via Adaptive Color Debiasing and Saturation Rectification

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A small adaptive color transform wrapped around any enhancement backbone removes the embedded color casts behind low-light color error.

desk verdict A well-scoped color correction module with a promising mechanism and solid ablations, but the headline numbers are undercut by uncontrolled baselines and dataset-specific post-processing. read the letter →

arxiv 2608.10512 v1 pith:ZYCMANIG submitted 2026-08-11 cs.CV

Zhichen Yang , Rui Xu , Yuzhen Niu , Fusheng Li , Hui Da , Ri Cheng This is my paper · ORCID
classification cs.CV
keywords low-lightimageenhancementcolor-faithfulrestorationadaptivecylindricalcolortransformAdaLABspacedebiasingsaturationrectificationout-of-gamutlightnesscompensationbias
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

The paper argues that the color casts and patchy saturation visible in enhanced low-light photos are not side effects of brightness recovery: they come from a color bias embedded in the raw image, which survives any change of color space and is then amplified by enhancement. CAGE wraps an existing enhancement backbone with two transforms built on a LAB-like space whose parameters adapt to each image: the forward transform shifts each pixel's chroma against the learned bias direction and rescales it, so the backbone receives a debiased input, and the inverse transform reverts the scaling while converting colors that would fall outside the display gamut into extra lightness instead of clipping them. Tested on three different backbones across LOLv1, LOLv2, SDSD, and SID, the wrapper reports consistent gains in color fidelity, saturation naturalness, and full-reference quality with only 0.07M added parameters and negligible FLOPs. If the paper is right, faithful color restoration is largely a wrapper problem: a cheap, backbone-agnostic correction layer, rather than a new enhancement architecture.

What carries the argument

The load-bearing mechanism is the pair AdaLAB and AdaCCT. AdaLAB is a cylindrical, image-adaptive form of CIELab that keeps the L axis and the two opponent a*b* axes but reorganizes the chromatic plane per image. AdaCCT is the transform that performs the reorganization, driven by three groups of image-adaptive parameters predicted from a 128×128 downsampled input: lightness sensitivity vertices (adaptive, monotonically increasing partitions of the lightness axis), chroma-scaling intensities (a learnable 1D map along lightness with exponentially decaying global interaction across all vertices), and a single two-dimensional hue-shift direction whose per-vertex magnitudes make the shift lightness-aware through linear interpolation. The forward transform computes $\tilde u_l = u_l - s_l d_l$ with a similarity weight $s_l \in [\delta_1, \delta_2]$ and then $\hat u_l = c_l \tilde u_l$; the inverse transform divides by the scaling ($\tilde u_h = \hat u_h / (c_h + \epsilon)$) and replaces gamut clipping with out-of-gamut lightness compensation, $l_h = \hat l_h + \gamma \| u_c - \tilde u_h \|_2$, converting unrepresentable chroma into a lightness gain. The transform is deliberately asymmetric: the inverse does not undo the debiasing, and the training target in AdaLAB space applies the same chroma scaling but omits the hue offset, so the backbone learns to enhance a debiased, well-scaled representation instead of inheriting the embedded cast.

What would settle it

Run CAGE on a mixed-illumination test set, for example a room lit by warm tungsten on one side and cool daylight on the other, and compare each output pixel with ground truth in a*b* coordinates: if spatially separated regions of nearly equal lightness show hue residuals with opposite signs or significantly different mean directions, the shared-direction debiasing is wrong exactly where the paper concedes it could fail.

Watch

Extended reading notes

Core claim

The central claim is that lightness–chrominance decoupling alone cannot remove the chromatic disturbance embedded in a low-light image: the disturbance propagates through any enhancement pipeline and reappears as global color bias together with local under- and over-saturation. CIELab offers a coherent basis for organizing this disturbance, but a reversible transform merely re-parameterizes it. CAGE therefore builds an explicitly asymmetric transform, AdaCCT, with a forward pass that suppresses the disturbance and an inverse pass that repairs its aftermath. The forward pass predicts, from a downsampled input, a shared two-dimensional hue-shift direction whose magnitude varies with lightness, shifts each pixel's chroma along that direction by a similarity-weighted amount, then rescales chroma by a lightness-dependent intensity learned with global interaction across lightness intervals. The inverse pass reverts only the scaling, not the debiasing, and handles out-of-gamut colors by turning the unrepresentable chroma surplus into a lightness gain rather than clipping it. The reported experiments on six benchmarks, three backbone families, no-reference datasets, and a human rating study support that this wrapper improves color fidelity and overall quality at negligible computational cost.

Load-bearing premise

The load-bearing premise is that within any one image the low-light color cast points in a single shared hue direction, with only its strength varying as brightness varies, so a single shift direction plus lightness-dependent magnitude can model and remove the bias; if a scene mixes different light sources and regions of similar brightness are cast in different directions, the debiasing can be wrong and residual local color bias persists.

Editorial extensions

If this is right

  • Color-faithful enhancement becomes a wrapper problem: any existing low-light backbone can be retrofitted with the forward and inverse transforms, and the reported gains come from color modeling rather than added capacity (0.07M parameters, under 0.01 GFLOPs).
  • The consistent gains across a Retinex-based transformer, an RGB-space network, and an HSV-inspired color-space network imply the embedded-bias mechanism is general to low-light imaging and not an artifact of one architecture.
  • Because the inverse transform turns out-of-gamut chroma into lightness gain instead of clipping, strong enhancement no longer has to trade saturated colors against washed-out ones; the reported reduction in over- and under-saturation is the direct consequence.
  • The reported no-reference improvements (BRISQUE and NIQE reductions in most of the 15 backbone-dataset combinations) imply the color correction carries over to real unpaired photos, which is the setting where such a module would actually be deployed.

Reading between the lines

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

  • Boundary prediction: the shared-direction premise implies the method will struggle exactly where the authors place its stated boundary — scenes with mixed light sources, where regions of similar brightness are cast in different directions; a spatially-varying hue-shift predictor is the natural next step, though it would trade the global-coherence prior that keeps the module at 0.07M parameters.
  • Out-of-gamut lightness compensation is a transferable principle: any operation that pushes colors outside a display gamut, such as tone mapping, color grading, or gamut mapping, could convert chroma surplus into lightness instead of clipping, and that claim could be tested independently on a gamut-mapping benchmark.
  • The paper's ablation splits color failure into two mechanisms — global bias removed by hue shifting and saturation abnormality fixed by chroma scaling — and reports that removing hue-shift hurts PSNR most while removing chroma-scaling hurts LPIPS most; that split could be reused as a diagnostic for attributing color problems in any given enhancement dataset.
  • A testable extension is to push the debiasing upstream into the RAW domain, where the color cast originates in the sensor pipeline; doing so would show whether the learned shift direction can correct bias before demosaicing and how OOGLC interacts with RAW-level denoising.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 CAGE, a plug-and-play color-correction wrapper for low-light image enhancement. It introduces AdaLAB, an image-adaptive variant of the CIELab space in which low-light color bias is modeled as a lightness-dependent hue-directional shift plus chroma scaling, and AdaCCT, a forward/inverse transform pair. The forward transform debiases and rescales chroma before the enhancement backbone operates; the inverse transform reverts chroma scaling and maps out-of-gamut chroma surplus to lightness compensation. The authors integrate CAGE into Retinexformer, DarkIR, and HVI-CIDNet and report consistent PSNR/SSIM/LPIPS gains on LOLv1, LOLv2, SDSD, and SID, plus no-reference gains on DICM/LIME/MEF/NPE/VV, ablation studies, and a human perceptual study.

Significance. If the evaluation is properly controlled, the paper's contribution is a low-cost, backbone-agnostic color-correction module with a clearly described mechanism. Strengths include the small overhead (about 0.07M additional parameters and less than 0.01 GFLOPs), the breadth of benchmarks (six paired and five no-reference datasets), ablations that isolate hue shift, chroma scaling, lightness sensitivity, color-space choice, and gamut handling, and the release of code. The central empirical claim currently rests on a comparison protocol that does not fully isolate CAGE from retraining and post-hoc adjustment effects, so the significance is conditional on the control experiments requested below.

major comments (4)
  1. [Sec. 4.1, Tables 1–2] The comparison protocol does not isolate CAGE. The text states that backbone networks are retrained for fair comparison, but no retrained baseline without CAGE is reported; the '+Ours' rows are compared against numbers whose provenance (published versus retrained) is unspecified. Because retraining recipes and seeds can shift PSNR/SSIM by the same order as some of the claimed gains, the reported deltas may partly reflect training configuration rather than AdaLAB/AdaCCT. Please report, for each backbone and dataset, a same-protocol retrained baseline without CAGE and the '+Ours' result initialized or trained under identical conditions.
  2. [Appendix B.1, Eq. (20)] Dataset-specific post-hoc adjustments and GT-mean rescaling are applied to CAGE outputs, but it is not stated whether they are applied identically to the baseline outputs. On LOLv1, CAGE uses alpha_l=1.3 and GT-mean rescaling (Eq. 20); on LOLv2-real, it uses alpha_l=1.1 and alpha_c=0.8. If the baselines are not adjusted in the same way, some of the reported gains (e.g., +2.21 dB on LOLv2-real for Retinexformer) could arise from these evaluation-side operations rather than from the transform. Please either apply the same adjustments to all compared methods or ablate the adjustments to show that CAGE's gains persist without them.
  3. [Sec. 3.4, Eq. (11)] The AdaLAB supervision target is constructed by applying the input-predicted chroma-scaling intensity to the ground truth. This makes part of the supervision self-referential: the model can reduce the AdaLAB loss by predicting a scaling that aligns the target to its own representation, rather than learning a genuinely corrective scaling. The authors should justify this design or ablate it against a fixed target (e.g., GT-derived scaling or no scaling in the AdaLAB loss), and should report whether the RGB-space loss alone preserves the observed gains.
  4. [Sec. 3.2, Appendix D] The method assumes a single image-level hue-shift direction whose magnitude varies with lightness, and Appendix D concedes that mixed illumination can violate this assumption, leaving residual local color bias. This limitation is acknowledged, but its impact is not quantified. To bound the central claim, please report results on mixed-illumination subsets or on the most challenging SDSD/SID scenes, and clarify whether the gains are concentrated in scenes with coherent global color bias.
minor comments (5)
  1. [Tables 1–2, LPIPS notation] The notation '+0.065' under LPIPS is confusing because lower LPIPS is better; the caption says '+(-) denotes improvement (reduction)', but a reader cannot tell whether '+0.065' means a 0.065 reduction (good) or an increase. Please use a direction-consistent delta convention for each metric.
  2. [Abstract, Fig. 8] The phrase 'cylindrical adaptive LAB color space' is slightly misleading because AdaLAB is displayed in Cartesian coordinates in Fig. 8; the cylindrical aspect lives in the parameterization of the transform rather than in the color-space coordinates. Consider rewording for clarity.
  3. [Fig. 3 caption] The term 'GT-mean Brighten Image' is not defined in the caption; please define it or point the reader to Appendix B.1 where the GT-mean operation is explained.
  4. [Sec. 4.4, Gamut Harmonization paragraph] The text says 'As shown in Table 8' when referring to the gamut-harmonization ablation, but the table displayed in the main text is Table 5, and the appendix also contains a Table 8. Please renumber to avoid ambiguity.
  5. [Appendix C, Human Subjective Evaluation] The human study is a useful addition, but no statistical significance test is reported for the mean-score differences. Please state whether the observed improvements over the three baselines are statistically significant.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CAGE is an empirical, end-to-end trained color transform; the auxiliary AdaLAB target is anchored by the RGB reconstruction loss, and self-citations are not load-bearing.

full rationale

CAGE does not claim a first-principles derivation. Its forward and inverse transforms are learned components trained end-to-end with a reconstruction loss in both AdaLAB and RGB space (Eq. 11). The only self-referential element is the AdaLAB target y_AdaLAB, which applies the input-predicted chroma scaling to the ground truth (Sec. 3.4). This is not circular because the same predicted scaling is divided out in the inverse transform (Eq. 8), so the RGB loss against the true ground truth anchors the final output; the AdaLAB loss is a consistency constraint rather than a substitute for ground-truth supervision. Self-citations [31, 44, 45] appear in related work and baseline comparisons and carry no load-bearing argument. The dataset-specific alpha_l/alpha_c settings and GT-mean rescaling (Appendix B.1) are evaluation-protocol choices, not derivation steps; they could affect comparison fairness but do not make the method's output equivalent to its input. Appendix D's acknowledged limitation on mixed illumination is an assumption about the target domain, not a circularity.

Assumptions & free parameters 6 free parameters · 5 assumptions · 2 invented entities

The central claim depends on a modest set of hand-set hyperparameters and dataset-specific output gains, plus a strong domain assumption about global coherence of color bias. The learned network parameters themselves are numerous but are the normal fitted weights of an empirical method. No new physical entity is introduced.

free parameters (6)
  • lightness interval count p = 32
    Selected by ablation on LOLv2-real (Fig. 11); values below or above reduce PSNR and SSIM.
  • similarity-weight bounds delta1, delta2 = 0.2, 1.0
    Fixed constants in Eq. (6b) controlling the range of the similarity-weighted shift; chosen by hand.
  • lightness compensation ratio gamma = 1.0
    Hand-set in Eq. (9b); no ablation over gamma is reported.
  • loss balance lambda = 1.0
    Hand-set balance between AdaLAB and RGB reconstruction losses in Eq. (11).
  • output saturation and brightness gains alpha_c, alpha_l = LOLv1: alpha_l=1.3, alpha_c=1.0; LOLv2-Real: alpha_l=1.1, alpha_c=0.8; default 1.0/1.0
    Dataset-specific adjustments in Appendix B.1 that directly modify final saturation and brightness before LAB-to-RGB conversion.
  • global-interaction temperature tau = learned
    Learnable scalar in Eq. (4a) controlling exponential decay across interval indices.
assumptions (5)
  • domain assumption Low-light color bias is globally coherent in direction and varies mainly with lightness.
    Used in Sec. 3.2 to predict a single hue-shift direction d with lightness-dependent magnitudes; the authors acknowledge in Appendix D that mixed illumination can violate this.
  • domain assumption CIELab provides a perceptually coherent basis for reorganizing low-light chromatic distributions.
    Central motivation in Secs. 1 and 3; the choice of LAB over RGB, YUV, and HVI is evaluated empirically but assumed to be the right coordinate system.
  • domain assumption The feasible chroma range in LAB is non-monotonic along the lightness axis, so chroma scaling should be non-monotonic.
    Invoked in Sec. 3.2 and Appendix B.4 to justify abandoning a monotonic 1D LUT for the global-interaction vertex model.
  • domain assumption Paired ground-truth images are valid color-faithful targets.
    All losses and metrics in Eq. (11) and Tables 1 and 2 assume the reference images define the desired color; dataset biases would propagate into the model.
  • domain assumption Existing enhancement backbones can be wrapped with forward and inverse color transforms without architectural changes.
    Appendix A.2 assumes Retinexformer, DarkIR, and HVI-CIDNet remain effective when their inputs and outputs are moved to AdaLAB.
invented entities (2)
  • AdaLAB color space
    purpose: Image-specific cylindrical LAB representation with lightness-dependent hue-shift and chroma-scaling parameters for decoupled color correction.
    Proposed by the paper as the working space for enhancement; no external falsifiable prediction, its value is supported only by in-paper experiments.
  • AdaCCT transform
    purpose: Adaptive forward and inverse cylindrical color transform performing debiasing before and saturation rectification after backbone enhancement.
    A computational construct whose claimed benefit is demonstrated only through the paper's own benchmarks and ablations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Color-Faithful Low-Light Image Enhancement via Adaptive Color Debiasing and Saturation Rectification." pith.science (2026). https://pith.science/paper/ZYCMANIG

@misc{pith2026260810512,
  author       = {Pith},
  title        = {Pith review of: Towards Color-Faithful Low-Light Image Enhancement via Adaptive Color Debiasing and Saturation Rectification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZYCMANIG}},
  note         = {Machine review of arXiv:2608.10512}
}
read the original abstract

Low-light imaging often introduces color bias caused by the low signal-to-noise ratio and the image formation process. Although recent low-light image enhancement methods have achieved strong brightness recovery, faithful color restoration remains challenging, manifesting as overall color bias together with local under- and over-saturation. To address this issue, we propose CAGE, a cylindrical color correction framework with adaptive color debiasing and gamut-harmonized saturation rectification for color-faithful low-light image enhancement. We first introduce AdaLAB, a cylindrical adaptive LAB color space that provides a decoupled and image-specific basis for uniform color correction. Building on this color space, we further develop AdaCCT, an adaptive cylindrical color transform with forward and inverse transforms for the conversion between RGB and AdaLAB color space, as well as necessary color debiasing and saturation rectification. The forward transform suppresses embedded color bias before backbone enhancement by reorganizing the chromatic distribution through chromatic-plane shifting and scaling, while the inverse transform achieves faithful saturation rectification through out-of-gamut lightness compensation. Extensive experiments on multiple benchmarks show that CAGE achieves more faithful color restoration, specifically reduces color bias and saturation abnormality, and delivers better overall visual quality across different low-light enhancement backbones. The code is available at https://yangzhichen763.github.io/CAGE/.

Figures

Figures reproduced from arXiv: 2608.10512 by the authors.

Figure 1
Figure 1. Motivation of our method. (a→b) Direct brightening makes the color shift pattern inside low-light images more visible. (a→c) Baseline enhancement improves brightness but still leaves a large chromatic gap from the well-lit image, with noticeable color bias, local under- and over-saturation. (a→d) Our method improves color faithfulness by adaptive color debiasing before backbone enhancement and gamut-harmonized satur… view at source ↗
Figure 2
Figure 2. Overview of our proposed framework with a plug-and-play adaptive cylindrical color transform built on the AdaLAB [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Visualization of low-light color shift patterns. The [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (16 more)
Figure 4
Figure 4. Figure 4: Pipeline of the forward and inverse transforms of the proposed AdaCCT. (a) The forward transform maps the input [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Comparison of Out-of-Gamut Handling Strategies. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Qualitative comparison of representative methods and baseline methods with and without our proposed CAGE on [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Visual comparison of different color spaces for LLIE. [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: LAB serves as a more coherent basis for color-faithful low-light image enhancement. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Detailed pipeline of the lightness-aware interpola [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Integration strategies of CAGE in three repre [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Ablation of the interval number 𝑝 [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]
Figure 13
Figure 13. Figure 13: Distribution of subjective scores for representative [PITH_FULL_IMAGE:figures/full_fig_p016_13.png]
Figure 12
Figure 12. Figure 12: Interface used for human subjective evaluation. [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]
Figure 14
Figure 14. Figure 14: Visual examples on the DICM, LIME, and MEF datasets among LIME [ [PITH_FULL_IMAGE:figures/full_fig_p017_14.png]
Figure 15
Figure 15. Figure 15: Visual examples on the NPE and VV datasets among UHDFour [ [PITH_FULL_IMAGE:figures/full_fig_p018_15.png]
Figure 16
Figure 16. Figure 16: Visual examples on the LOLv1 dataset among RUAS [ [PITH_FULL_IMAGE:figures/full_fig_p018_16.png]
Figure 17
Figure 17. Figure 17: Visual examples on the LOLv2-real dataset among KinD [ [PITH_FULL_IMAGE:figures/full_fig_p019_17.png]
Figure 18
Figure 18. Figure 18: Visual examples for low-light image enhancement on the LOLv2-synthetic dataset among RUAS, KinD [ [PITH_FULL_IMAGE:figures/full_fig_p020_18.png]
Figure 19
Figure 19. Figure 19: Visual examples for low-light image enhancement on the SDSD-indoor, SDSD-outdoor, and SID datasets among [PITH_FULL_IMAGE:figures/full_fig_p021_19.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

68 extracted references · 59 canonical work pages

  1. [1]

    Mahmoud Afifi, Luxi Zhao, Abhijith Punnappurath, Mohamed A Abdelsalam, Ran Zhang, and Michael S Brown. 2025. Time-aware auto white balance in mobile photography. InProceedings of the IEEE/CVF International Conference on Computer Vision. 5038–5047

    work page 2025

  2. [2]

    Yousef Atoum, Mao Ye, Liu Ren, Ying Tai, and Xiaoming Liu. 2020. Color- wise attention network for low-light image enhancement. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. 506–507

    work page 2020

  3. [3]

    Jiesong Bai, Yuhao Yin, Qiyuan He, Yuanxian Li, and Xiaofeng Zhang. 2024. RetinexMamba: Retinex-based Mamba for low-light image enhancement. In Proceedings of the International Conference on Neural Information Processing. Springer, 427–442

    work page 2024

  4. [4]

    Alexandru Brateanu, Raul Balmez, Adrian Avram, Ciprian Orhei, and Cosmin Ancuti. 2025. LYT-NET: Lightweight YUV transformer-based network for low- light image enhancement.IEEE Signal Processing Letters32 (2025), 2065–2069

    work page 2025

  5. [5]

    Yuanhao Cai, Hao Bian, Jing Lin, Haoqian Wang, Radu Timofte, and Yulun Zhang. 2023. Retinexformer: One-stage Retinex-based transformer for low-light image enhancement. InProceedings of the IEEE/CVF International Conference on Computer Vision. 12504–12513

    work page 2023

  6. [6]

    Luyang Cao, Han Xu, Jian Zhang, Lei Qi, Jiayi Ma, Yinghuan Shi, and Yang Gao. 2025. Towards perfection: Building inter-component mutual correction for Retinex-based low-light image enhancement. InProceedings of the ACM International Conference on Multimedia. 9549–9558

    work page 2025

  7. [7]

    Chen Chen, Qifeng Chen, Jia Xu, and Vladlen Koltun. 2018. Learning to see in the dark. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 3291–3300

    work page 2018

  8. [8]

    Ziteng Cui, Kunchang Li, Lin Gu, Shenghan Su, Peng Gao, Zhengkai Jiang, Yu Qiao, and Tatsuya Harada. 2022. You only need 90k parameters to adapt light: A light weight transformer for image enhancement and exposure correc- tion, In Proceedings of the British Machine Vision Conference.arXiv preprint arXiv:2205.14871

    arXiv 2022

Show all 68 references
  1. [9]

    Dan Ding, Feng Shi, and Ye Li. 2025. Low-illumination color imaging: Progress and challenges.Optics and Laser Technology184 (2025), 112553

    2025

  2. [10]

    Daniel Feijoo, Juan C Benito, Alvaro Garcia, and Marcos V Conde. 2025. DarkIR: Robust low-light image restoration. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 10879–10889

    2025

  3. [11]

    Theo Gevers, Arjan Gijsenij, Joost Van de Weijer, and Jan-Mark Geusebroek. 2012. Color in computer vision: Fundamentals and applications. John Wiley & Sons

    2012

  4. [12]

    Chunle Guo, Chongyi Li, Jichang Guo, Chen Change Loy, Junhui Hou, Sam Kwong, and Runmin Cong. 2020. Zero-reference deep curve estimation for low- light image enhancement. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 1780–1789

    2020

  5. [13]

    Hang Guo, Jinmin Li, Tao Dai, Zhihao Ouyang, Xudong Ren, and Shu-Tao Xia

  6. [14]

    Xiaojie Guo and Qiming Hu. 2023. Low-light image enhancement via breaking down the darkness.International Journal of Computer Vision131 (2023), 48–66

    2023

  7. [15]

    Xiaojie Guo, Yu Li, and Haibin Ling. 2016. LIME: Low-light image enhancement via illumination map estimation.IEEE Transactions on Image Processing26 (2016), 982–993

    2016

  8. [16]

    Quan Huynh-Thu and Mohammed Ghanbari. 2008. Scope of validity of PSNR in image/video quality assessment.Electronics Letters44 (2008), 800–801

    2008

  9. [17]

    2019.Colorimetry — Part 4: CIE 1976 L*a*b* Colour Space

    International Commission on Illumination and International Organization for Standardization. 2019.Colorimetry — Part 4: CIE 1976 L*a*b* Colour Space. ISO and CIE

    2019

  10. [18]

    Hai Jiang, Ao Luo, Xiaohong Liu, Songchen Han, and Shuaicheng Liu. 2024. Ligh- tenDiffusion: Unsupervised low-light image enhancement with latent-Retinex diffusion models. InProceedings of the European Conference on Computer Vision. 161–179

    2024

  11. [19]

    Yifan Jiang, Xinyu Gong, Ding Liu, Yu Cheng, Chen Fang, Xiaohui Shen, Jian- chao Yang, Pan Zhou, and Zhangyang Wang. 2021. EnlightenGAN: Deep light enhancement without paired supervision.IEEE Transactions on Image Processing 30 (2021), 2340–2349

    2021

  12. [20]

    Hanul Kim, Su-Min Choi, Chang-Su Kim, and Yeong Jun Koh. 2021. Represen- tative color transform for image enhancement. InProceedings of the IEEE/CVF International Conference on Computer Vision. 4459–4468

    2021

  13. [21]

    Wontae Kim, Keuntek Lee, and Nam Ik Cho. 2025. Lightweight and fast real-time image enhancement via decomposition of the spatial-aware lookup tables. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 11895– 11905

    2025

  14. [22]

    Chulwoo Lee, Chul Lee, and Chang-Su Kim. 2013. Contrast enhancement based on layered difference representation of 2D histograms.IEEE Transactions on Image Processing22 (2013), 5372–5384

    2013

  15. [23]

    Chongyi Li, Chunle Guo, Linghao Han, Jun Jiang, Ming-Ming Cheng, Jinwei Gu, and Chen Change Loy. 2021. Low-light image and video enhancement using deep learning: A survey.IEEE transactions on pattern analysis and machine intelligence 44, 12 (2021), 9396–9416

    2021

  16. [24]

    Chongyi Li, Chunle Guo, and Chen Change Loy. 2021. Learning to enhance low-light image via zero-reference deep curve estimation.IEEE Transactions on Pattern Analysis and Machine Intelligence44 (2021), 4225–4238

    2021

  17. [25]

    Chongyi Li, Chun-Le Guo, Man Zhou, Zhexin Liang, Shangchen Zhou, Ruicheng Feng, and Chen Change Loy. 2023. Embedding fourier for ultra-high-definition low-light image enhancement. InProceedings of the International Conference on Learning Representations

    2023

  18. [26]

    Risheng Liu, Long Ma, Jiaao Zhang, Xin Fan, and Zhongxuan Luo. 2021. Retinex- inspired unrolling with cooperative prior architecture search for low-light image enhancement. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 10561–10570

    2021

  19. [27]

    Junyu Lou, Xiaorui Zhao, Kexuan Shi, and Shuhang Gu. 2025. Learning pixel- adaptive multi-layer perceptrons for real-time image enhancement. InProceedings of the IEEE/CVF International Conference on Computer Vision. 14095–14105

    2025

  20. [28]

    Kede Ma, Kai Zeng, and Zhou Wang. 2015. Perceptual quality assessment for multi-exposure image fusion.IEEE Transactions on Image Processing24 (2015), 3345–3356

    2015

  21. [29]

    Anish Mittal, Anush Krishna Moorthy, and Alan Conrad Bovik. 2012. No- reference image quality assessment in the spatial domain.IEEE Transactions on Image Processing21 (2012), 4695–4708

    2012

  22. [30]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C Bovik. 2012. Making a “completely blind” image quality analyzer.IEEE Signal Processing Letters20 (2012), 209–212

    2012

  23. [31]

    Yuzhen Niu, Fusheng Li, Yuezhou Li, Siling Chen, and Yuzhong Chen. 2025. Adaptive luminance enhancement and high-fidelity color correction for low-light image enhancement.IEEE Transactions on Computational Imaging11 (2025), 732–747

    2025

  24. [32]

    David Serrano-Lozano, Luis Herranz, Michael S Brown, and Javier Vazquez- Corral. 2024. NamedCurves: Learned image enhancement via color naming. In Proceedings of the European Conference on Computer Vision. 92–108

    2024

  25. [33]

    Vassilios Vonikakis, Rigas Kouskouridas, and Antonios Gasteratos. 2018. On the evaluation of illumination compensation algorithms.Multimedia Tools and Applications77 (2018), 9211–9231

    2018

  26. [34]

    Chenxi Wang and Zhi Jin. 2023. Brighten-and-colorize: A decoupled network for customized low-light image enhancement. InProceedings of the ACM International Conference on Multimedia. 8356–8366

    2023

  27. [35]

    Chenxi Wang, Hongjun Wu, and Zhi Jin. 2023. FourLLIE: Boosting low-light image enhancement by fourier frequency information. InProceedings of the ACM International Conference on Multimedia. 7459–7469

    2023

  28. [36]

    Ruixing Wang, Xiaogang Xu, Chi-Wing Fu, Jiangbo Lu, Bei Yu, and Jiaya Jia. 2021. Seeing dynamic scene in the dark: A high-quality video dataset with mechatronic alignment. InProceedings of the IEEE/CVF International Conference on Computer Vision. 9700–9709

    2021

  29. [37]

    Shuhang Wang, Jin Zheng, Hai-Miao Hu, and Bo Li. 2013. Naturalness preserved enhancement algorithm for non-uniform illumination images.IEEE Transactions on Image Processing22 (2013), 3538–3548

    2013

  30. [38]

    Tao Wang, Kaihao Zhang, Tianrun Shen, Wenhan Luo, Bjorn Stenger, and Tong Lu. 2023. Ultra-high-definition low-light image enhancement: A benchmark and transformer-based method. InProceedings of the AAAI Conference on Artificial Intelligence. 2654–2662

    2023

  31. [39]

    Wenjing Wang, Huan Yang, Jianlong Fu, and Jiaying Liu. 2024. Zero-reference low- light enhancement via physical quadruple priors. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 26057–26066

    2024

  32. [40]

    Yufei Wang, Renjie Wan, Wenhan Yang, Haoliang Li, Lap-Pui Chau, and Alex Kot. 2022. Low-light image enhancement with normalizing flow. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 36. 2604–2612

    2022

  33. [41]

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. 2004. Image quality assessment: From error visibility to structural similarity.IEEE Transactions on Image Processing13 (2004), 600–612

    2004

  34. [42]

    Chen Wei, Wenjing Wang, Wenhan Yang, and Jiaying Liu. 2018. Deep Retinex decomposition for low-light enhancement. InProceedings of the British Machine Vision Conference

    2018

  35. [43]

    Wenhui Wu, Jian Weng, Pingping Zhang, Xu Wang, Wenhan Yang, and Jianmin Jiang. 2022. URetinex-Net: Retinex-based deep unfolding network for low-light image enhancement. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 5901–5910. MM ’26, Nove...

    2022

  36. [44]

    Rui Xu, Yuezhou Li, Yuzhen Niu, Huangbiao Xu, Yuzhong Chen, and Tiesong Zhao. 2024. Bilateral interaction for local-global collaborative perception in low-light image enhancement.IEEE Transactions on Multimedia26 (2024), 10792– 10804

    2024

  37. [45]

    Rui Xu, Yuzhen Niu, Yuezhou Li, Huangbiao Xu, Wenxi Liu, and Yuzhong Chen

  38. [46]

    Xiaogang Xu, Ruixing Wang, Chi-Wing Fu, and Jiaya Jia. 2022. SNR-aware low- light image enhancement. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 17714–17724

    2022

  39. [47]

    Qingsen Yan, Yixu Feng, Cheng Zhang, Guansong Pang, Kangbiao Shi, Peng Wu, Wei Dong, Jinqiu Sun, and Yanning Zhang. 2025. HVI: A new color space for low-light image enhancement. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 5678–5687

    2025

  40. [48]

    Qingsen Yan, Kangbiao Shi, Yixu Feng, Tao Hu, Peng Wu, Guansong Pang, and Yanning Zhang. 2025. HVI-CIDNet+: Beyond extreme darkness for low-light image enhancement.arXiv preprint arXiv:2507.06814(2025)

    2025 arXiv

  41. [49]

    Canqian Yang, Meiguang Jin, Xu Jia, Yi Xu, and Ying Chen. 2022. AdaInt: Learning adaptive intervals for 3D lookup tables on real-time image enhancement. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 17522–17531

    2022

  42. [50]

    Canqian Yang, Meiguang Jin, Yi Xu, Rui Zhang, Ying Chen, and Huaida Liu

  43. [51]

    Wenhan Yang, Shiqi Wang, Yuming Fang, Yue Wang, and Jiaying Liu. 2020. From fidelity to perceptual quality: A semi-supervised approach for low-light image enhancement. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 3063–3072

    2020

  44. [52]

    Wenhan Yang, Wenjing Wang, Haofeng Huang, Shiqi Wang, and Jiaying Liu

  45. [53]

    Xunpeng Yi, Han Xu, Hao Zhang, Linfeng Tang, and Jiayi Ma. 2025. Diff- Retinex++: Retinex-driven reinforced diffusion model for low-light image en- hancement.IEEE Transactions on Pattern Analysis and Machine Intelligence(2025), 6823–6841

    2025

  46. [54]

    Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shah- baz Khan, and Ming-Hsuan Yang. 2022. Restormer: Efficient transformer for high-resolution image restoration. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 5728–5739

    2022

  47. [55]

    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. InProceedings of the European Conference on Computer Vision. 492–511

    2020

  48. [56]

    Hui Zeng, Jianrui Cai, Lida Li, Zisheng Cao, and Lei Zhang. 2020. Learning image-adaptive 3d lookup tables for high performance photo enhancement in real-time.IEEE Transactions on Pattern Analysis and Machine Intelligence44 (2020), 2058–2073

    2020

  49. [57]

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang

  50. [58]

    Tongshun Zhang, Pingping Liu, Yubing Lu, Mengen Cai, Zijian Zhang, Zhe Zhang, and Qiuzhan Zhou. 2025. CWNet: Causal wavelet network for low-light image enhancement. InProceedings of the IEEE/CVF International Conference on Computer Vision. 8789–8799

    2025

  51. [59]

    Yu Zhang, Xiaoguang Di, Bin Zhang, Ruihang Ji, and Chunhui Wang. 2021. Better than reference in low-light image enhancement: Conditional re-enhancement network.IEEE Transactions on Image Processing31 (2021), 759–772

    2021

  52. [60]

    Yonghua Zhang, Xiaojie Guo, Jiayi Ma, Wei Liu, and Jiawan Zhang. 2021. Beyond brightening low-light images.International Journal of Computer Vision129 (2021), 1013–1037

    2021

  53. [61]

    Yonghua Zhang, Jiawan Zhang, and Xiaojie Guo. 2019. Kindling the darkness: A practical low-light image enhancer. InProceedings of the ACM International Conference on Multimedia. 1632–1640

    2019

  54. [62]

    Han Zhou, Wei Dong, Xiaohong Liu, Shuaicheng Liu, Xiongkuo Min, Guangtao Zhai, and Jun Chen. 2024. GLARE: Low light image enhancement via generative latent feature based codebook retrieval. InProceedings of the European Conference on Computer Vision. 36–54

    2024

  55. [63]

    Shangchen Zhou, Chongyi Li, and Chen Change Loy. 2022. LEDNet: Joint low- light enhancement and deblurring in the dark. InProceedings of the European Conference on Computer Vision. 573–589. Towards Color-Faithful Low-Light Image Enhancement via Adaptive Color Debiasing and Sat...

    2022

  56. [2018]

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

    The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 586–595

  57. [2021]

    Sparse gradient regularized deep Retinex network for robust low-light image enhancement.IEEE Transactions on Image Processing30 (2021), 2072–2086

    2021

  58. [2022]

    InProceedings of the European Conference on Computer Vision

    SepLUT: Separable image-adaptive lookup tables for real-time image enhancement. InProceedings of the European Conference on Computer Vision. 201–217

  59. [2024]

    InProceedings of the European Conference on Computer Vision

    MambaIR: A simple baseline for image restoration with state-space model. InProceedings of the European Conference on Computer Vision. 222–241

  60. [2025]

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

    URWKV: Unified RWKV model with multi-state perspective for low-light image restoration. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 21267–21276

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.