Pith. sign in

REVIEW 2 major objections 6 minor 58 references

Lightweight and Fast Real-time Image Enhancement via Decomposition of the Spatial-aware Lookup Tables

T0 review · 2 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Image-adaptive color lookup tables — the workhorse of real-time photo enhancement — are so redundant that a full 3D table can be replaced by three weighted 2D tables of rank eight, an 88 percent parameter cut with no quality loss.

desk verdict Solid LUT-efficiency paper: real contribution, small quality margins, and one under-tested decomposition assumption that a good referee can nail down. read the letter →

arxiv 2508.16121 v1 pith:JAAYJB3Q submitted 2025-08-22 eess.IV cs.CV

Wontae Kim , Keuntek Lee , Nam Ik Cho This is my paper · ORCID
classification eess.IVcs.CV
keywords imageenhancement3Dlookuptable2DLUTdecompositionsingularvaluebilateralgridcache-efficientinferencereal-timephotoretouchspatial-awarecolortransform
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

Real-time photo enhancement often works by building a 3D lookup table, a cube of color-to-color mappings, tailored to each image. This paper establishes that those per-image tables are overwhelmingly redundant: fewer than one in ten vertices is ever looked up, and accesses pile up near the diagonal. It replaces the 3D table with a weighted sum of three 2D tables, one for each color-channel pair, and stores each 2D table as its singular-value decomposition keeping only eight singular values. The result is a spatial-aware enhancer with 160.5K parameters that reaches 25.76 dB PSNR at 480p and processes a 4K frame in 1.38 ms — about a third of the parameters and a third of the 4K runtime of the previous best spatial-aware method. The paper also shows that the high-resolution bottleneck is memory traffic rather than arithmetic, and fuses the spatial-slicing and table-lookup steps so that few large intermediate tensors are ever written.

What carries the argument

The SVD-compressed 2D LUT is the key mechanism. Each output channel's transform is a weighted sum of three 2D tables T_rg, T_rb, T_gb (33×33 each), and a small generator produces each table as three SVD factors — U (33×8), S (8 values), Vᵀ (8×33) — so 1,608 stored values per channel replace a 33³ cube of 35,937. A toy experiment justifies rank 8: truncating a pretrained full-rank 2D table keeps PSNR, while the same truncation on bilateral grids hurts, so grids are reduced from 3D to 2D but not SVD-compressed. The second mechanism is cache-effective fusion: grid slicing and LUT transform run in one fused pass, never writing high-resolution intermediate feature maps. This drops 4K runtime from

What would settle it

On a dataset of edits with strong hue rotation or selective channel mixing — for instance cross-processed film looks where one output channel depends on all three inputs at once — fit a full-rank per-image 3D LUT and measure the reconstruction error of the rank-8 pairwise approximation, or train the proposed model and compare PSNR against a full 3D-LUT model at the same parameter budget. If the gap widens on those edits, the low-rank pairwise assumption is content-dependent. The paper's own toy experiment — truncate a pretrained full-rank 2D LUT and check PSNR — supplies the template.

Watch

Extended reading notes

Core claim

The central claim: a per-image 3D color transform does not need to be a 3D object. On FiveK, a predicted 33³ LUT references fewer than 10% of its vertices, with accesses concentrated along the diagonal; 1D tables saturate, 2D tables sit at a comfortable utilization. So each output channel's 3D table is replaced by a weighted sum of three 2D tables, one per channel pair (rg, rb, gb), with per-image scalar weights, and each 2D table stored as U·S·Vᵀ with eight singular values — a rank justified by truncating a pretrained full-rank table with negligible PSNR loss. Bilateral grids supply spatial information, reduced 3D-to-2D but kept full-rank since they degrade under truncation. The reported re

Load-bearing premise

The load-bearing premise is that a single image's color transformation is so redundant that three pairwise 2D tables of rank eight capture everything a full 3D table does — every genuine three-channel interaction at a vertex is assumed negligible or representable by per-image scalar weights.

Editorial extensions

If this is right

  • Real-time spatial-aware enhancement at 4K becomes practical on a single GPU: 1.38 ms per frame versus 3.64 ms for the prior best spatial-aware method, with the authors noting the gap should widen on weaker hardware.
  • The parameter budget drops to 160.5K — about one third of SABLUT's 463.7K and roughly 1/28 of the 4.5M of the original spatial-aware 3D LUT — while PSNR improves on FiveK at both 480p (25.76 dB) and 4K (25.69 dB).
  • The decomposition transfers across tasks: results are reported for photo retouch on FiveK and PPR10K, tone mapping on FiveK, and SDR-to-HDR video conversion on HDRTV1K, all at real-time speed.
  • Because the 4K runtime fix is about memory traffic rather than arithmetic, other LUT-based pipelines (super-resolution, HDR reconstruction) can adopt the fused slicing-and-transform pass without changing their accuracy, as the paper itself suggests as future work.

Reading between the lines

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

  • The 10% utilization and diagonal-concentration figures imply that photo-editing color transforms are nearly per-channel (1D-like) with cross-channel coupling as a small correction. If so, other color-mapping architectures — 1D LUT chains, polynomial grades, small MLPs — might compress far below their reported sizes using the same rank analysis; the paper does not test this.
  • The rank-8 choice is justified entirely on the photo-retouch task. A stress test on edits with hue rotation or cross-processing, where one output channel's value genuinely depends on all three inputs at once, would reveal whether the pairwise/rank-8 structure is content-dependent.
  • The bilateral grid's failure to survive SVD truncation is itself informative: spatial maps are high-dimensional while color maps are low-rank. A cheaper intermediate design, such as keeping grids 2D but compressing the spatial feature tensor after slicing, might capture the same spatial information at lower cost.
  • The cache-fusion result is separable from the decomposition: it suggests that any multi-stage pixel-wise operator that materializes high-resolution intermediates — unsharp-mask-style pipelines, guided filters, two-pass LUTs — is memory-bound at 4K and could fuse stages the same way, independent of the SVD idea.
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

2 major / 6 minor

Summary. The manuscript proposes SVDLUT, a real-time image-enhancement network that replaces the image-adaptive 3D LUT and 3D bilateral grid of SABLUT with weighted sums of three pairwise 2D LUTs / 2D bilateral grids, compresses the 2D LUTs by SVD to rank 8, and merges slicing and LUT transformation into a single cache-efficient pass. On FiveK, PPR10K, and HDRTV1K, the reported model uses 160.5K parameters, reaches 25.76/25.69 dB PSNR at 480p/4K on FiveK, and cuts 4K runtime from 3.64 ms (SABLUT) to 1.38 ms while maintaining PSNR/SSIM/ΔE. The main claims are (i) a 3D LUT can be replaced by a linear combination of 2D LUTs with negligible loss, (ii) SVD rank-8 is sufficient for the LUT, and (iii) fused slicing/LUT transform reduces high-resolution runtime.

Significance. If the claims hold, the contribution is useful: it reduces model size by roughly 3x over SABLUT, keeps quality within 0.03–0.10 dB on the main benchmark, and removes the 4K runtime penalty of spatial-aware LUTs. The paper is reproducible: code is released, and the ablations are controlled in several dimensions (LUT dimension, grid dimension, SVD rank, component contribution, module runtime). The toy SVD-truncation experiment (Fig. 4a) is independent evidence that the learned 2D LUTs are compressible, and Table 6 isolates the contribution of the cache-efficient fusion. The main weakness is that the pairwise 2D decomposition itself is only tested implicitly; no block-level approximation error is reported. The practical margins over SABLUT are small but consistent, and the runtime gain is resolution-dependent.

major comments (2)
  1. [Sec. 3.2.1, Eq. (4)] The central decomposition claim—that an image-adaptive 3D LUT can be replaced by a scalar-weighted sum of three pairwise 2D LUTs—is not directly quantified. Table 1 is an end-to-end ablation in which both the LUT and bilateral-grid branches are retrained; Fig. 3 reports vertex-access statistics, which speak to sparsity, not to the approximation error of Eq. (4). Fig. 4a truncates a pre-trained 2D LUT by SVD and therefore tests only rank reduction, not the 3D-to-2D additive structure. Because Eq. (4) has no three-way interaction term, transforms with nonzero mixed third derivatives cannot be represented exactly; whether this matters for real enhancements is exactly the point that needs direct measurement. Please add an experiment that takes a pre-trained full 3D-LUT model, replaces the 3D LUTs with the best-fit weighted 2D-LUT decomposition while keeping the rest frozen (or retraining onl
  2. [Sec. 3.2.2 / Sec. 5] The choice of rank N_s=8 and the 2D structure are validated on the same datasets/architecture on which the final model is trained and evaluated. The toy truncation in Fig. 4a is from a single pre-trained model, and the final model only ever produces rank-8 2D LUTs, so the good end-to-end results are partly a consequence of the chosen inductive bias. This does not invalidate the results, but it means the reported gains are tied to a task-specific hyperparameter. The concluding sentence in Sec. 5 that the decomposition 'can be applied to other frameworks' goes beyond the evidence. Please add a rank/dimension sweep on a held-out task (HDRTV1K is already in the supplement) or explicitly frame the rank and the 2D decomposition as task-specific architecture choices.
minor comments (6)
  1. [Supplementary, Eq. (16)] In the weighted sum for the 2D LUT transform, the term w^c_rb · φ^c_rb appears twice; the third term should be w^c_gb · φ^c_gb.
  2. [Table 2] In the original-structure column, the row 'LUT/weight Gen' lists 0.29 ms at 480p and 0.05 ms at 4K. This ordering appears to be a typo and should be corrected.
  3. [Tables 3 and 7] At 480p the proposed method is 0.17 ms slower than SABLUT on V100 (1.37 vs 1.20 ms) and 0.19 ms slower on GTX 1660 SUPER (1.64 vs 1.45 ms). The runtime advantage is specific to high resolution; the abstract and introduction should state this more precisely rather than implying a general runtime improvement.
  4. [Sec. 1] The '88% reduction' in parameters is not tied to a clear baseline. Specify whether it refers to the LUT table only, at which D_t, and relative to the 3D LUT before or after the 2D decomposition.
  5. [Sec. 3.2.1, Eq. (7)] The channel indexing c'_k and k' = mod(k,3) is confusing. Define c' consistently and make explicit the distinction between the input color channel used for the lookup and the output spatial-feature channel produced by the slicing operation.
  6. [Sec. 4.5.1, Table 5] On PPR10K-b, the proposed method is slightly worse than SABLUT on ΔE_ab (7.23 vs 7.19) and ΔE_HC (4.68 vs 4.66). The sentence that the method 'outperforms other methods' should be qualified to reflect this mixed result.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the pairwise-additive 2D LUT decomposition and the rank-8 SVD choice are empirically grounded in direct ablations and SVD-truncation experiments, not derived from the final model's own output.

full rationale

The central claim -- that an image-adaptive 3D LUT can be replaced by a weighted sum of three 2D LUTs (Eq. 4) -- is presented as a hypothesis and tested by the ablation in Table 1, where 3D and 2D LUT/grid variants of the same SABLUT-style baseline are trained and compared. This is an independent end-to-end comparison, not a tautology: the 2D-LUT model could have failed. The rank-8 SVD choice is similarly supported by the toy experiment in Fig. 4a, which applies SVD truncation to a full-rank pre-trained 2D LUT and measures the resulting PSNR before the final low-rank model is introduced; the final trained model (Fig. 4c) is reported as corroboration, not as the sole evidence. Final metrics are measured on held-out test splits of FiveK, PPR10K, and HDRTV1K against external baselines. The paper does cite the authors' own SABLUT [22] for the backbone, loss hyperparameters, and as a baseline, but those citations are code-reproduced/peer-reviewed and are not used as an unchallengeable uniqueness theorem or to forbid alternative decompositions; they do not supply the low-rank or pairwise-additive content. The skeptic concern that the 3D-to-2D additive approximation error of Eq. (4) is not directly quantified on real image-adaptive 3D LUTs is a legitimate validation gap, but it is a correctness/portability risk, not a circular reduction of the paper's predictions to its inputs.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The paper does not introduce new physical entities, forces, or measured quantities. Its inventions are architectural: a 2D LUT family, SVD factors U, S, V as network outputs, low-rank degrees of freedom, and a fused kernel. These are treated as free parameters and domain assumptions above rather than as entities, since they are not independently evidenced objects in the world.

free parameters (6)
  • SVD rank N_s = 8
    Chosen from Fig. 4c/4d experiments; performance saturates around eight singular values for the LUT. Tuned on the FiveK task (supplementary says Dt=33 was chosen after 25.54 dB at Dt=17, and eight singular values after observing the curve).
  • LUT resolution D_t = 33
    Supplementary: 17 vertices gives 25.54 dB, 33 gives 25.76 dB; 33 is adopted. Selection is based on the target benchmark.
  • Bilateral grid dimension D_s = 17
    Set to 17, K=6 for fair comparison with SABLUT; influenced by baseline rather than independently justified.
  • Number of bilateral grids K = 6
    K=6 inherited from SABLUT for a fair spatial-feature comparison; no ablation is given.
  • Generator hidden sizes M_t, M_sw, M_tw = 8
    Bottleneck MLP widths are 'experimentally selected' (supplementary), i.e., tuned on the validation task.
  • Decomposition weights w_rg, w_rb, w_gb and bias b = per-image, predicted by H_tw/H_sw
    These are learned per-image blending coefficients that carry the channel-correlation information lost in the 2D factorization; their capacity and role are not separately ablated.
assumptions (4)
  • domain assumption Image-adaptive LUTs and bilateral grids are redundant so a 3D table can be replaced by 2D tables with a weighted sum and a low-rank factor
    Established by the usage-rate statistics in Fig. 3 and the truncation experiments in Fig. 4; it is an empirical property of natural images and learned mappings, not a theorem.
  • domain assumption Bilinear interpolation of 2D tables is a sufficient replacement for trilinear interpolation of a 3D table
    The 2D tables are evaluated on channel pairs (r,g), (r,b), (g,b), so the cross-channel interaction is handled only through the weighted sum. This is assumed to have enough capacity.
  • domain assumption The 4K runtime bottleneck is memory traffic from intermediate high-resolution tensors, and fusing kernels removes it
    Used to justify the cache-effective structure in Sec. 3.2.3 and Table 2; supported only by end-to-end runtime numbers, no memory-bandwidth profiling is given.
  • standard math Standard deep-learning machinery (Adam, MSE plus CIE94 and LPIPS losses, LeakyReLU/instance-norm backbone) trains the SVD components end to end
    The training procedure in Sec. 4.2 assumes standard optimizers and losses; no novel optimization claim is made.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lightweight and Fast Real-time Image Enhancement via Decomposition of the Spatial-aware Lookup Tables." pith.science (2026). https://pith.science/paper/JAAYJB3Q

@misc{pith2026250816121,
  author       = {Pith},
  title        = {Pith review of: Lightweight and Fast Real-time Image Enhancement via Decomposition of the Spatial-aware Lookup Tables},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JAAYJB3Q}},
  note         = {Machine review of arXiv:2508.16121}
}
read the original abstract

The image enhancement methods based on 3D lookup tables (3D LUTs) efficiently reduce both model size and runtime by interpolating pre-calculated values at the vertices. However, the 3D LUT methods have a limitation due to their lack of spatial information, as they convert color values on a point-by-point basis. Although spatial-aware 3D LUT methods address this limitation, they introduce additional modules that require a substantial number of parameters, leading to increased runtime as image resolution increases. To address this issue, we propose a method for generating image-adaptive LUTs by focusing on the redundant parts of the tables. Our efficient framework decomposes a 3D LUT into a linear sum of low-dimensional LUTs and employs singular value decomposition (SVD). Furthermore, we enhance the modules for spatial feature fusion to be more cache-efficient. Extensive experimental results demonstrate that our model effectively decreases both the number of parameters and runtime while maintaining spatial awareness and performance.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 54 canonical work pages

  1. [1]

    A two-stage deep network for high dynamic range image reconstruction

    SM A Sharif, Rizwan Ali Naqvi, Mithun Biswas, and Sungjun Kim. A two-stage deep network for high dynamic range image reconstruction. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 550–559, 2021. 1

    work page 2021

  2. [2]

    Deep joint deinterlacing and de- noising for single shot dual-iso hdr reconstruction

    Ahmet O ˘guz Aky ¨uz et al. Deep joint deinterlacing and de- noising for single shot dual-iso hdr reconstruction. IEEE Transactions on Image Processing, 29:7511–7524, 2020. 1

    work page 2020

  3. [3]

    Learning photographic global tonal adjustment with a database of input/output image pairs

    Vladimir Bychkovsky, Sylvain Paris, Eric Chan, and Fr ´edo Durand. Learning photographic global tonal adjustment with a database of input/output image pairs. InCVPR 2011, pages 97–104. IEEE, 2011. 1, 3, 4, 5, 6, 7, 8, 12, 15, 16

    work page 2011

  4. [4]

    Learning a deep single image contrast enhancer from multi-exposure images

    Jianrui Cai, Shuhang Gu, and Lei Zhang. Learning a deep single image contrast enhancer from multi-exposure images. IEEE Transactions on Image Processing, 27(4):2049–2062,

    work page

  5. [5]

    Learning to see in the dark

    Chen Chen, Qifeng Chen, Jia Xu, and Vladlen Koltun. Learning to see in the dark. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 3291–3300, 2018. 1

    work page 2018

  6. [6]

    Real-time edge-aware image processing with the bilateral grid

    Jiawen Chen, Sylvain Paris, and Fr ´edo Durand. Real-time edge-aware image processing with the bilateral grid. ACM Transactions on Graphics (TOG), 26(3):103–es, 2007. 4

    work page 2007

  7. [7]

    Bilateral guided upsampling

    Jiawen Chen, Andrew Adams, Neal Wadhwa, and Samuel W Hasinoff. Bilateral guided upsampling. ACM Transactions on Graphics (TOG), 35(6):1–8, 2016

    work page 2016

  8. [8]

    Hdrunet: Single image hdr reconstruc- tion with denoising and dequantization

    Xiangyu Chen, Yihao Liu, Zhengwen Zhang, Yu Qiao, and Chao Dong. Hdrunet: Single image hdr reconstruc- tion with denoising and dequantization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 354–363, 2021. 1

    work page 2021

Show all 58 references
  1. [9]

    Deep photo enhancer: Unpaired learning for im- age enhancement from photographs with gans

    Yu-Sheng Chen, Yu-Ching Wang, Man-Hsin Kao, and Yung- Yu Chuang. Deep photo enhancer: Unpaired learning for im- age enhancement from photographs with gans. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 6306–6314, 2018. 1, 6, 7

    2018

  2. [10]

    Nilut: Conditional neural implicit 3d lookup tables for image enhancement

    Marcos V Conde, Javier Vazquez-Corral, Michael S Brown, and Radu Timofte. Nilut: Conditional neural implicit 3d lookup tables for image enhancement. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 1371– 1379, 2024. 7, 8

    2024

  3. [11]

    Aesthetic- driven image enhancement by adversarial learning

    Yubin Deng, Chen Change Loy, and Xiaoou Tang. Aesthetic- driven image enhancement by adversarial learning. In Pro- ceedings of the 26th ACM international conference on Mul- timedia, pages 870–878, 2018. 1

    2018

  4. [12]

    Deep bilateral learning for real- time image enhancement

    Micha ¨el Gharbi, Jiawen Chen, Jonathan T Barron, Samuel W Hasinoff, and Fr´edo Durand. Deep bilateral learning for real- time image enhancement. ACM Transactions on Graphics (TOG), 36(4):1–12, 2017. 6, 7, 15

    2017

  5. [13]

    Con- ditional sequential modulation for efficient global image re- touching

    Jingwen He, Yihao Liu, Yu Qiao, and Chao Dong. Con- ditional sequential modulation for efficient global image re- touching. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceed- ings, Part XIII 16, pages 679–695. Springer, 2020. 6, 7, 15

    2020

  6. [14]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 6

    2016

  7. [15]

    Exposure: A white-box photo post-processing framework

    Yuanming Hu, Hao He, Chenxi Xu, Baoyuan Wang, and Stephen Lin. Exposure: A white-box photo post-processing framework. ACM Transactions on Graphics (TOG) , 37(2): 1–17, 2018

    2018

  8. [16]

    Hdr-nerf: High dynamic range neu- ral radiance fields

    Xin Huang, Qi Zhang, Ying Feng, Hongdong Li, Xuan Wang, and Qing Wang. Hdr-nerf: High dynamic range neu- ral radiance fields. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 18398–18408, 2022. 1

    2022

  9. [17]

    Dslr-quality photos on mobile devices with deep convolutional networks

    Andrey Ignatov, Nikolay Kobyshev, Radu Timofte, Kenneth Vanhoey, and Luc Van Gool. Dslr-quality photos on mobile devices with deep convolutional networks. InProceedings of the IEEE international conference on computer vision, pages 3277–3285, 2017

    2017

  10. [18]

    Enlightengan: Deep light enhancement without paired supervision

    Yifan Jiang, Xinyu Gong, Ding Liu, Yu Cheng, Chen Fang, Xiaohui Shen, Jianchao Yang, Pan Zhou, and Zhangyang Wang. Enlightengan: Deep light enhancement without paired supervision. IEEE transactions on image processing, 30:2340–2349, 2021. 1

    2021

  11. [19]

    Practical single-image super-resolution using look-up table

    Younghyun Jo and Seon Joo Kim. Practical single-image super-resolution using look-up table. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 691–700, 2021. 1

    2021

  12. [20]

    Pienet: Personalized image enhancement network

    Han-Ul Kim, Young Jun Koh, and Chang-Su Kim. Pienet: Personalized image enhancement network. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXX 16 , pages 374–390. Springer, 2020. 1

    2020

  13. [21]

    Jsi-gan: Gan- based joint super-resolution and inverse tone-mapping with pixel-wise task-specific filters for uhd hdr video

    Soo Ye Kim, Jihyong Oh, and Munchurl Kim. Jsi-gan: Gan- based joint super-resolution and inverse tone-mapping with pixel-wise task-specific filters for uhd hdr video. InProceed- ings of the AAAI Conference on Artificial Intelligence, pages 11287–11295, 2020. 1

    2020

  14. [22]

    Image-adaptive 3d lookup tables for real-time image enhancement with bilateral grids

    Wontae Kim and Nam Ik Cho. Image-adaptive 3d lookup tables for real-time image enhancement with bilateral grids. In European Conference on Computer Vision, pages 91–108,

  15. [23]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 ,

    arXiv

  16. [24]

    Imagenet classification with deep convolutional neural net- works

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural net- works. Advances in neural information processing systems , 25, 2012. 5 9

    2012

  17. [25]

    Single-image hdr reconstruction by multi-exposure generation

    Phuoc-Hieu Le, Quynh Le, Rang Nguyen, and Binh-Son Hua. Single-image hdr reconstruction by multi-exposure generation. In Proceedings of the IEEE/CVF Winter Confer- ence on Applications of Computer Vision, pages 4063–4072,

  18. [26]

    Mulut: Cooperating multiple look-up tables for efficient im- age super-resolution

    Jiacheng Li, Chang Chen, Zhen Cheng, and Zhiwei Xiong. Mulut: Cooperating multiple look-up tables for efficient im- age super-resolution. In European conference on computer vision, pages 238–256. Springer, 2022. 1

    2022

  19. [27]

    Look-up table compression for efficient image restoration

    Yinglong Li, Jiacheng Li, and Zhiwei Xiong. Look-up table compression for efficient image restoration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 26016–26025, 2024. 1, 2

    2024

  20. [28]

    Ppr10k: A large-scale portrait photo retouch- ing dataset with human-region mask and group-level consis- tency

    Jie Liang, Hui Zeng, Miaomiao Cui, Xuansong Xie, and Lei Zhang. Ppr10k: A large-scale portrait photo retouch- ing dataset with human-region mask and group-level consis- tency. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, pages 653–661...

    2021

  21. [29]

    4d lut: learnable context-aware 4d lookup table for image enhancement

    Chengxu Liu, Huan Yang, Jianlong Fu, and Xueming Qian. 4d lut: learnable context-aware 4d lookup table for image enhancement. IEEE Transactions on Image Processing, 32: 4742–4756, 2023. 1, 2

    2023

  22. [30]

    Reconstructed convolution module based look-up tables for efficient image super-resolution

    Guandu Liu, Yukang Ding, Mading Li, Ming Sun, Xing Wen, and Bin Wang. Reconstructed convolution module based look-up tables for efficient image super-resolution. In Proceedings of the IEEE/CVF international conference on computer vision, pages 12217–12226, 2023. 1

    2023

  23. [31]

    Single-image hdr reconstruction by learning to reverse the camera pipeline

    Yu-Lun Liu, Wei-Sheng Lai, Yu-Sheng Chen, Yi-Lung Kao, Ming-Hsuan Yang, Yung-Yu Chuang, and Jia-Bin Huang. Single-image hdr reconstruction by learning to reverse the camera pipeline. In Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition , pages...

    2020

  24. [32]

    Learn- ing series-parallel lookup tables for efficient image super- resolution

    Cheng Ma, Jingyi Zhang, Jie Zhou, and Jiwen Lu. Learn- ing series-parallel lookup tables for efficient image super- resolution. In European Conference on Computer Vision , pages 305–321. Springer, 2022. 1

    2022

  25. [33]

    Deeplpf: Deep local para- metric filters for image enhancement

    Sean Moran, Pierre Marza, Steven McDonagh, Sarah Parisot, and Gregory Slabaugh. Deeplpf: Deep local para- metric filters for image enhancement. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12826–12835, 2020. 1, 7

    2020

  26. [34]

    Distort-and-recover: Color enhancement using deep reinforcement learning

    Jongchan Park, Joon-Young Lee, Donggeun Yoo, and In So Kweon. Distort-and-recover: Color enhancement using deep reinforcement learning. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 5928–5936, 2018

    2018

  27. [35]

    U- net: Convolutional networks for biomedical image segmen- tation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- net: Convolutional networks for biomedical image segmen- tation. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Pa...

    2015

  28. [36]

    A survey on 3d-lut performance in 10-bit and 12-bit hdr bt

    Jean D Vandenberg and Stefano Andriani. A survey on 3d-lut performance in 10-bit and 12-bit hdr bt. 2100 pq. In SMPTE 2018, pages 1–19. SMPTE, 2018. 3

    2018

  29. [37]

    Underexposed photo enhance- ment using deep illumination estimation

    Ruixing Wang, Qing Zhang, Chi-Wing Fu, Xiaoyong Shen, Wei-Shi Zheng, and Jiaya Jia. Underexposed photo enhance- ment using deep illumination estimation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6849–6857, 2019. 6, 7

    2019

  30. [38]

    Real-time image en- hancer via learnable spatial-aware 3d lookup tables

    Tao Wang, Yong Li, Jingyang Peng, Yipeng Ma, Xian Wang, Fenglong Song, and Youliang Yan. Real-time image en- hancer via learnable spatial-aware 3d lookup tables. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 2471–2480, 2021. 1, 2, 5, ...

    2021

  31. [39]

    Image quality assessment: from error visibility to structural similarity

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Si- moncelli. Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing, 13(4):600–612, 2004. 6

    2004

  32. [40]

    Deep retinex decomposition for low-light enhancement

    Chen Wei, Wenjing Wang, Wenhan Yang, and Jiaying Liu. Deep retinex decomposition for low-light enhancement. arXiv preprint arXiv:1808.04560, 2018. 1

    2018 arXiv

  33. [41]

    Bilateral grid learning for stereo matching networks

    Bin Xu, Yuhua Xu, Xiaoli Yang, Wei Jia, and Yulan Guo. Bilateral grid learning for stereo matching networks. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12497–12506, 2021

    2021

  34. [42]

    Automatic photo adjustment using deep neu- ral networks

    Zhicheng Yan, Hao Zhang, Baoyuan Wang, Sylvain Paris, and Yizhou Yu. Automatic photo adjustment using deep neu- ral networks. ACM Transactions on Graphics (TOG), 35(2): 1–15, 2016. 1

    2016

  35. [43]

    Adaint: Learning adaptive intervals for 3d lookup tables on real-time image enhancement

    Canqian Yang, Meiguang Jin, Xu Jia, Yi Xu, and Ying Chen. 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 (CVPR), pages 17522–17531, 2022. 1, 2, 5, 6, ...

    2022

  36. [44]

    Seplut: Separable image-adaptive lookup tables for real-time image enhancement

    Canqian Yang, Meiguang Jin, Yi Xu, Rui Zhang, Ying Chen, and Huaida Liu. Seplut: Separable image-adaptive lookup tables for real-time image enhancement. In Euro- pean Conference on Computer Vision (ECCV) , pages 201–

  37. [45]

    Taming lookup tables for efficient image retouching

    Sidi Yang, Binxiao Huang, Mingdeng Cao, Yatai Ji, Hanzhong Guo, Ngai Wong, and Yujiu Yang. Taming lookup tables for efficient image retouching. In European Confer- ence on Computer Vision, pages 144–159. Springer, 2024. 7, 8

    2024

  38. [46]

    Learning image-adaptive 3d lookup tables for high perfor- mance photo enhancement in real-time

    Hui Zeng, Jianrui Cai, Lida Li, Zisheng Cao, and Lei Zhang. Learning image-adaptive 3d lookup tables for high perfor- mance photo enhancement in real-time. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) , 44 (4):2058–2073, 2022. 1, 2, 5, 6, 7, 8, 12, 15

  39. [47]

    Clut-net: Learning adaptively compressed representations of 3dluts for lightweight image enhancement

    Fengyi Zhang, Hui Zeng, Tianjun Zhang, and Lin Zhang. Clut-net: Learning adaptively compressed representations of 3dluts for lightweight image enhancement. In Proceedings of the 30th ACM International Conference on Multimedia , pages 6493–6501, 2022. 1, 2

    2022

  40. [48]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 5 10

    2018

  41. [49]

    Kindling the darkness: A practical low-light image enhancer

    Yonghua Zhang, Jiawan Zhang, and Xiaojie Guo. Kindling the darkness: A practical low-light image enhancer. In Pro- ceedings of the 27th ACM international conference on mul- timedia, pages 1632–1640, 2019. 1

    2019

  42. [50]

    Ultra-high- definition image dehazing via multi-guided bilateral learn- ing

    Zhuoran Zheng, Wenqi Ren, Xiaochun Cao, Xiaobin Hu, Tao Wang, Fenglong Song, and Xiuyi Jia. Ultra-high- definition image dehazing via multi-guided bilateral learn- ing. In 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 16180–16189. IEEE, 2021

    2021

  43. [51]

    Ultra-high-definition image hdr reconstruc- tion via collaborative bilateral learning

    Zhuoran Zheng, Wenqi Ren, Xiaochun Cao, Tao Wang, and Xiuyi Jia. Ultra-high-definition image hdr reconstruc- tion via collaborative bilateral learning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 4449–4458, 2021

    2021

  44. [52]

    4k-resolution photo exposure correction at 125 fps with˜ 8k parameters

    Yijie Zhou, Chao Li, Jin Liang, Tianyi Xu, Xin Liu, and Jun Xu. 4k-resolution photo exposure correction at 125 fps with˜ 8k parameters. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision , pages 1587–1597, 2024

    2024

  45. [53]

    Hdr-vdp-2: A calibrated visual metric for visibility and quality predictions in all luminance conditions

    Rafat Mantiuk, Kil Joong Kim, Allan G Rempel, and Wolf- gang Heidrich. Hdr-vdp-2: A calibrated visual metric for visibility and quality predictions in all luminance conditions. ACM Transactions on graphics (TOG), 30(4):1–14, 2011. 15

    2011

  46. [54]

    A new journey from sdrtv to hdrtv

    Xiangyu Chen, Zhengwen Zhang, Jimmy S Ren, Lynhoo Tian, Yu Qiao, and Chao Dong. A new journey from sdrtv to hdrtv. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision , pages 4500–4509, 2021. 15, 18

    2021

  47. [55]

    Objective metric for the assessment of the poten- tial visibility of colour differences in television

    ITU-R. Objective metric for the assessment of the poten- tial visibility of colour differences in television. ITU-R Rec BT.2124-0, 2019. 15 11 Lightweight and Fast Real-time Image Enhancement via Decomposition of the Spatial-aware Lookup Tables Supplementary Material In this s...

    2019

  48. [100]

    The overall tendency is similar to the results of the analysis for FiveK in the main paper

    The occurrence statistics are estimated by counting the number of accesses for each vertex. The overall tendency is similar to the results of the analysis for FiveK in the main paper. Fig. 7a presents LUT utilization rate of PPR10K. The rate of 3D LUT is very low, and 1D LUT i...

    2020

  49. [217]

    2, 3, 5, 6, 7, 12

    Springer, 2022. 2, 3, 5, 6, 7, 12

    2022

  50. [2024]

    1, 2, 3, 4, 5, 6, 7, 12, 15

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.