Pith. sign in

REVIEW 3 major objections 6 minor 88 references

Video Color Grading via Look-Up Table Generation

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read The paper claims a diffusion model conditioned on the difference of high-level reference and input features can generate a 3D look-up table that transfers a reference's artistic grade to an entire video while preserving structure…

desk verdict Clever diffusion-based LUT generation, but the evaluation leans on a synthetic benchmark that rewards LUT inversion; the mood-transfer claim rests on a thin user study. read the letter →

arxiv 2508.00548 v1 pith:CTHZUHQD submitted 2025-08-01 cs.CV

classification cs.CV
keywords videocolorgradinglook-uptablediffusionmodelreference-basedretouchingtemporalconsistencytext-promptCLIPfeatures
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 tries to turn video color grading into an explicit, reusable operation: instead of regenerating frames in latent space, a diffusion model predicts a single 3D look-up table that shifts the input video's colors toward a reference frame's artistic intent. The authors argue this captures high-level attributes such as look, mood, and emotion, not just color distributions, and that applying the same table to every frame preserves structural detail and temporal consistency. If the claim holds, video grading becomes accessible to non-specialists: pick a reference image or video, get a consistent grade, and optionally refine it with a text prompt such as 'increase contrast.' The paper supports the claim with quantitative comparisons over seven baselines and user studies on mood and intent matching.

What carries the argument

The load-bearing mechanism is a pair of networks. A Grading Style Extraction network (GS-Extractor), initialized from a pretrained latent diffusion model, encodes the reference and source key frames into high-level features; its conditioning vector is the difference $C = G(z_{I'_{\hat n}}^0) - G(z_{I_{\hat m}}^0)$, which points from the input's style to the reference's style. A LUT diffusion model (L-Diffuser) takes that difference, along with an identity LUT, and denoises a reshaped LUT deviation $\Delta L \in \mathbb{R}^{16 \times 16 \times 16 \times 3}$; adding the identity LUT back yields the final 3D color mapping. Because the output is a global LUT rather than regenerated pixels, the original frame content is untouched, and one table applied to every frame enforces temporal consistency by construction.

What would settle it

Take input footage paired with a professionally graded reference from a different film, where no ground-truth LUT exists, and have colorists choose which output best matches the reference's stated intent. If the generated LUTs only recover the synthetic LUTs used in training and fail on genuine cross-film grading examples, the claim of high-level mood transfer would be refuted.

Watch

Extended reading notes

Core claim

The central discovery is that explicit generation of a 3D LUT through a denoising diffusion model, conditioned on the vector difference of learned grading-style features between reference and input frames, separates the artistic part of grading from the content part. The L-Diffuser is trained to output small perturbations of the identity LUT, so the final mapping is a smooth color transform by construction; this transform is then applied uniformly to all frames, making temporal coherence a structural property rather than a constraint. The paper reports higher PSNR, SSIM, and LPIPS than seven baselines on its movie-derived test set and on Adobe5k, lower artifact selection rates in a user study, and better user ratings for mood and intent matching, along with the shortest inference time among the compared methods.

Load-bearing premise

The load-bearing premise is that grading a video by applying a hand-picked LUT to create ground-truth pairs, and using another frame from the same movie as the reference, faithfully represents real artistic color grading; if it does not, the reported gains may largely reflect LUT inversion rather than transfer of mood and intent.

Editorial extensions

If this is right

  • Applying one generated LUT to every frame makes temporal flicker a non-issue by construction, with no optical-flow warping or per-frame consistency losses.
  • Because the pipeline recolors rather than regenerates frames, the high-frequency detail and structural content of the original footage survive the grade.
  • Inference is dominated by generating a single LUT per video (about 12 seconds for 480 frames in the paper's setup), which makes the approach practical for batch or interactive grading.
  • The text-prompt retouching stage gives users a second editing pass by retrieving a LUT description, and the paper's user study indicates participants preferred the modified outputs on average.
  • The method also applies to still-image retouching, as demonstrated on Adobe5k, suggesting the mechanism is not specific to video.

Reading between the lines

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

  • Because the main benchmark is built by inverting known LUTs on frames from the same movie, I expect the quantitative lead to shrink on cross-film, artist-authored references; the user study is the evidence that would carry the artistic claim.
  • The ablation showing similar results across several standard image embedding models for key-frame selection suggests frame matching is not the bottleneck, so a cheaper matcher could replace CLIP without hurting the grade.
  • The paper's stated limitation at scene changes points to a natural extension: segment the video into scenes and generate one LUT per segment, turning scene cuts from a failure mode into a scheduling problem.
  • Since the LUT is global, spatially-variant grading such as recoloring only faces or skies is out of reach unless LUT generation is conditioned on semantic maps.
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

3 major / 6 minor

Summary. The paper proposes a reference-based video color grading framework that generates an explicit 3D look-up table (LUT) via a diffusion model (L-Diffuser). The LUT is conditioned on the difference between high-level features of the reference and input frames, extracted by a trained GS-Extractor network, and is then applied uniformly across all video frames to ensure temporal consistency. The pipeline also includes CLIP-based key-frame selection and a text-prompt-driven user feedback stage for additional retouching. The method is evaluated on a Condensed Movie dataset and Adobe5k against seven baselines using PSNR, SSIM, LPIPS, BRISQUE, and blur metrics, plus two user studies for subjective qualities such as mood and intent.

Significance. The central idea of generating an explicit LUT with a diffusion model is a sensible way to combine global color transfer with structural fidelity and fast inference, and the reported inference time advantage is practically relevant. The paper ships code and includes ablations that support several design choices (Table 3). However, the quantitative evaluation's synthetic same-video LUT setup and the limited statistical power of the user study mean that the distinctive claim of transferring mood and emotion is not yet convincingly isolated from low-level color matching. If the evaluation is strengthened, this could be a solid contribution to the video color grading literature.

major comments (3)
  1. [§4.1, Table 1] The CondensedMovie test protocol applies one of 10 held-out LUTs to the test video and uses the last frame of that same video after LUT application as the reference. Because source, reference, and target are related by a single known, content-aligned color mapping, the large PSNR/LPIPS margins (e.g., 24.55 vs 21.41 PSNR) can be explained by LUT inversion rather than by generalized artistic grade transfer. The abstract and Section 1 frame the task as transferring mood and emotion, but this experiment does not directly measure that attribute. Please add a cross-content evaluation (e.g., reference frames from a different video with no shared LUT) and report the same metrics there, or explicitly reframe the CondensedMovie results as a LUT-inversion sanity check rather than evidence for artistic transfer.
  2. [§4.5, Table 4] The user study is the only evidence directly bearing on the mood/emotion claim, but it reports no significance tests, no inter-rater reliability, and no control condition. With 20 participants and 15 items, the observed differences (Ours 4.14 vs next-best 3.80 for Mood Matching; Ours 3.97 vs 3.76 for Intent Matching) may not be statistically robust, and participants' ratings could be driven by low-level histogram similarity rather than perceived mood or intent. Please add paired statistical tests with appropriate multiple-comparison correction, report effect sizes and confidence intervals, and include a simple color-matching baseline (or a histogram-matched condition) to show that the mood/intent judgments are not reducible to color statistics.
  3. [Tables 1–3] None of the quantitative tables report standard deviations or multiple-run results, so it is unclear whether small differences on BRISQUE and Blur, and even some PSNR/SSIM gaps, are reproducible. For example, in Table 1 the BRISQUE difference between Ours (40.23) and CCPL (40.40) is tiny. Please report the number of runs and error bars, or justify why a single run is sufficient for these stochastic methods.
minor comments (6)
  1. [§3.1] The paper states that 100 LUT bases are selected as 'distinctive' from the 400 LUTs of the Video Harmonization Dataset, but the selection criteria are not described; please clarify how distinctiveness was operationalized.
  2. [§4.5, first paragraph] The sentence 'Each study has 15 questionnaires with 20 participants' is ambiguous: it could mean 20 participants per questionnaire or 20 total per study. Please clarify the participant allocation.
  3. [Table 4] The Artifact column reports selection rates, but it is not clear how the dummy choices were used to filter invalid responses; please describe this procedure explicitly.
  4. [§3.2, Eqs. (3)–(5)] The notation for source, reference, and target frames is dense and could confuse readers; for instance, in Eq. (3) the denoising network uses z^{I_m}_0 as source and G(z^{I'_n}_0) as reference conditioning, while Eq. (5) uses C = G(reference) - G(source). Please spell out these roles in the text.
  5. [§5] The first sentence of Limitations & Future Work contains a typo: 'There is still rooms for improvements' should be 'There is still room for improvement.'
  6. [References] Reference [62] is a blog URL for the Parasite cinematography analysis; consider citing a more established or peer-reviewed source for this widely discussed claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the LUT prediction is trained on a 90/10 held-out split and the central derivation does not reduce to its inputs.

full rationale

The paper's claimed derivation chain is self-contained. GS-Extractor is trained with the denoising objective in Eq. (3), and L-Diffuser is trained separately with the diffusion objective in Eq. (5), conditioned on C = G(z_ref) - G(z_source). The test evaluation is not fitted to the test targets: the paper states 'We divide LUTs with 90 and 10 LUTs for training and test, respectively,' and the CondensedMovie benchmark applies 10 held-out LUTs to generate Input/GT pairs. The reference frame is the last frame of the same video after the same held-out LUT is applied, which makes the benchmark an example-based LUT-recovery task and weakens its external validity for artistic mood transfer, but it does not make the model's output equivalent to its training inputs by construction. The only author self-citations ([51], [63], [64]) are used for user-study protocol and future-work pointers, not as load-bearing justification of the core LUT-generation claim. No self-definitional equation, fitted-parameter-as-prediction, or imported uniqueness argument was found. The central claim therefore has independent empirical content, and the paper's main risk is benchmark validity rather than circularity.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central method rests on several domain assumptions about the dataset, the feature space, and the evaluation protocol. The hand-picked LUT subset and unspecified LUT augmentation scheme are free choices that influence generalization.

free parameters (3)
  • LUT base subset size = 100 selected from 400
    The paper hand-selects 100 distinctive LUTs from the Video Harmonization Dataset; the choice determines which color styles the model sees and is not derived from an objective criterion.
  • 3D LUT grid resolution = 16x16x16
    The diffusion model outputs a 16^3 LUT; this resolution is chosen without reported analysis of trade-offs between color precision and memory.
  • Diffusion inference steps = 25 DDIM steps
    Number of denoising steps at inference is chosen by the authors; no sensitivity analysis is reported.
assumptions (5)
  • domain assumption Applying a single global 3D LUT to all frames can represent the color grading relationship between reference and input
    The entire pipeline assumes a spatially uniform transform; the limitation section admits spatially-variant grading is unsupported.
  • domain assumption GS-Extractor features encode style and mood beyond color distribution
    L-Diffuser is conditioned on the difference of these features; if they encode content rather than mood, the model cannot transfer artistic intent.
  • domain assumption CLIP features are a valid basis for key-frame selection
    Key-frame selection uses cosine similarity of CLIP embeddings without validating that this aligns with color-grading-relevant attributes.
  • domain assumption LUTs from the Video Harmonization Dataset span the space of artistic color grading styles
    Training and test LUTs come from this database; no evidence is provided that they represent real cinematic looks beyond harmonization.
  • domain assumption Reference frames in the user study are representative and non-overlapping with training data
    Famous movie frames may appear in Condensed Movie Dataset; no leakage analysis is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Video Color Grading via Look-Up Table Generation." pith.science (2026). https://pith.science/paper/CTHZUHQD

@misc{pith2026250800548,
  author       = {Pith},
  title        = {Pith review of: Video Color Grading via Look-Up Table Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CTHZUHQD}},
  note         = {Machine review of arXiv:2508.00548}
}
read the original abstract

Different from color correction and transfer, color grading involves adjusting colors for artistic or storytelling purposes in a video, which is used to establish a specific look or mood. However, due to the complexity of the process and the need for specialized editing skills, video color grading remains primarily the domain of professional colorists. In this paper, we present a reference-based video color grading framework. Our key idea is explicitly generating a look-up table (LUT) for color attribute alignment between reference scenes and input video via a diffusion model. As a training objective, we enforce that high-level features of the reference scenes like look, mood, and emotion should be similar to that of the input video. Our LUT-based approach allows for color grading without any loss of structural details in the whole video frames as well as achieving fast inference. We further build a pipeline to incorporate a user-preference via text prompts for low-level feature enhancement such as contrast and brightness, etc. Experimental results, including extensive user studies, demonstrate the effectiveness of our approach for video color grading. Codes are publicly available at https://github.com/seunghyuns98/VideoColorGrading.

Figures

Figures reproduced from arXiv: 2508.00548 by the authors.

Figure 1
Figure 1. (Left) Difference between conventional style/color transfer and the color grading task. Conventional color/style transfer aims to [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An overview of our video color grading framework: (1) key-frame selection; (2) LUT generation; (3) user preference feedback [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Comparison results of analysis. LP and L2 denote perceptual loss and ℓ2 loss, respectively. (Please zoom-in) Method PSNR↑ SSIM↑ LPIPS↓ BRISQUE↓ Blur↓ Ours w/ perceptual loss 20.87 0.7560 0.2666 43.38 0.4141 Ours w/ ℓ2 Loss 22.17 0.7857 0.2773 44.10 0.4205 NLUT w/ Explicit LUT 13.97 0.5936 0.4679 41.02 0.4244 Ours + LUT Bases-based 20.50 0.7321 0.3562 46.12 0.4220 Ours + MLP-based 19.73 0.7219 0.4014 42.35 0.4344 Our… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Example questionnaires for User Study 1, and outcomes from ours and the comparison methods [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Examples of user feedback on initally graded image. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

88 extracted references · 66 canonical work pages

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ah- mad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 ,

  2. [2]

    Histogan: Controlling colors of gan-generated and real im- ages via color histograms

    Mahmoud Afifi, Marcus A Brubaker, and Michael S Brown. Histogan: Controlling colors of gan-generated and real im- ages via color histograms. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 7941–7950, 2021. 2, 5, 6

  3. [3]

    Ultrafast photorealistic style transfer via neural architecture search

    Jie An, Haoyi Xiong, Jun Huan, and Jiebo Luo. Ultrafast photorealistic style transfer via neural architecture search. In Proceedings of the AAAI Conference on Artificial Intel- ligence, pages 10443–10450, 2020. 2, 5

  4. [4]

    Artflow: Unbiased image style transfer via re- versible neural flows

    Jie An, Siyu Huang, Yibing Song, Dejing Dou, Wei Liu, and Jiebo Luo. Artflow: Unbiased image style transfer via re- versible neural flows. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 862–871, 2021. 2

  5. [5]

    Condensed movies: Story based retrieval with con- textual embeddings

    Max Bain, Arsha Nagrani, Andrew Brown, and Andrew Zis- serman. Condensed movies: Story based retrieval with con- textual embeddings. In Proceedings of the Asian Conference on Computer Vision, 2020. 3

  6. [6]

    Example-based video color grading

    Nicolas Bonneel, Kalyan Sunkavalli, Sylvain Paris, and Hanspeter Pfister. Example-based video color grading. ACM Trans. Graph., 32(4):39–1, 2013. 1

  7. [7]

    In- structpix2pix: Learning to follow image editing instructions

    Tim Brooks, Aleksander Holynski, and Alexei A Efros. In- structpix2pix: Learning to follow image editing instructions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18392–18402, 2023. 4

  8. [8]

    Magicdance: Realistic human dance video gen- eration with motions & facial expressions transfer

    Di Chang, Yichun Shi, Quankai Gao, Jessica Fu, Hongyi Xu, Guoxian Song, Qing Yan, Xiao Yang, and Mohammad So- leymani. Magicdance: Realistic human dance video gen- eration with motions & facial expressions transfer. arXiv preprint arXiv:2311.12052, 2023. 4

Show all 88 references
  1. [9]

    Coherent online video style transfer

    Dongdong Chen, Jing Liao, Lu Yuan, Nenghai Yu, and Gang Hua. Coherent online video style transfer. In Proceedings of the IEEE International Conference on Computer Vision , pages 1105–1114, 2017. 2

  2. [10]

    Stylebank: An explicit representation for neural im- age style transfer

    Dongdong Chen, Lu Yuan, Jing Liao, Nenghai Yu, and Gang Hua. Stylebank: An explicit representation for neural im- age style transfer. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1897–1906,

  3. [11]

    Nlut: Neural-based 3d lookup tables for video photorealistic style transfer.arXiv preprint arXiv:2303.09170, 2023

    Yaosen Chen, Han Yang, Yuexin Yang, Yuegen Liu, Wei Wang, Xuming Wen, and Chaoping Xie. Nlut: Neural-based 3d lookup tables for video photorealistic style transfer.arXiv preprint arXiv:2303.09170, 2023. 3, 5, 6

  4. [12]

    Style-aware normalized loss for improving arbitrary style transfer

    Jiaxin Cheng, Ayush Jaiswal, Yue Wu, Pradeep Natara- jan, and Prem Natarajan. Style-aware normalized loss for improving arbitrary style transfer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 134–143, 2021. 2

  5. [13]

    Pca-based knowledge distillation towards lightweight and content-style balanced photorealistic style transfer models

    Tai-Yin Chiu and Danna Gurari. Pca-based knowledge distillation towards lightweight and content-style balanced photorealistic style transfer models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7844–7853, 2022. 2

  6. [14]

    Photowct2: Compact autoencoder for photorealistic style transfer resulting from blockwise training and skip connections of high-frequency residuals

    Tai-Yin Chiu and Danna Gurari. Photowct2: Compact autoencoder for photorealistic style transfer resulting from blockwise training and skip connections of high-frequency residuals. In Proceedings of the IEEE/CVF winter confer- ence on applications of computer vision , pages 2868–2877,

  7. [15]

    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. 3

  8. [16]

    The blur effect: perception and estimation with a new no-reference perceptual blur metric

    Frederique Crete, Thierry Dolmiere, Patricia Ladret, and Marina Nicolas. The blur effect: perception and estimation with a new no-reference perceptual blur metric. InHuman vi- sion and electronic imaging XII, pages 196–206. SPIE, 2007. 5

  9. [17]

    Amazon mechanical turk: A research tool for organizations and information systems scholars

    Kevin Crowston. Amazon mechanical turk: A research tool for organizations and information systems scholars. InShap- ing the Future of ICT Research. Methods and Approaches: IFIP WG 8.2, Working Conference, Tampa, FL, USA, De- cember 13-14, 2012. Proceedings, pages 210–221. Springer,

  10. [18]

    Arbitrary video style transfer via multi-channel correlation

    Yingying Deng, Fan Tang, Weiming Dong, Haibin Huang, Chongyang Ma, and Changsheng Xu. Arbitrary video style transfer via multi-channel correlation. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 1210– 1217, 2021. 2

  11. [19]

    Stytr2: Im- age style transfer with transformers

    Yingying Deng, Fan Tang, Weiming Dong, Chongyang Ma, Xingjia Pan, Lei Wang, and Changsheng Xu. Stytr2: Im- age style transfer with transformers. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11326–11336, 2022. 2

  12. [20]

    Max Welling Diederik P. Kingma. Auto-encoding variational bayes. In The Twelfth International Conference on Learning Representations, 2014. 4

  13. [21]

    A learned representation for artistic style

    Vincent Dumoulin, Jonathon Shlens, and Manjunath Kud- lur. A learned representation for artistic style. arXiv preprint arXiv:1610.07629, 2016. 2

  14. [22]

    The 10 most iconic color grades in film

    Filmworks. The 10 most iconic color grades in film. https://filmworkz.com/the-10-most-iconic-color-grades-in- fi/, 2024. 7, 8

  15. [23]

    Fast video multi-style transfer

    Wei Gao, Yijun Li, Yihang Yin, and Ming-Hsuan Yang. Fast video multi-style transfer. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , pages 3222–3230, 2020. 2

  16. [24]

    Im- age style transfer using convolutional neural networks

    Leon A Gatys, Alexander S Ecker, and Matthias Bethge. Im- age style transfer using convolutional neural networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2414–2423, 2016. 2

  17. [25]

    Two birds, one stone: A unified framework for joint learning of image and video style transfers

    Bohai Gu, Heng Fan, and Libo Zhang. Two birds, one stone: A unified framework for joint learning of image and video style transfers. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pages 23545–23554,

  18. [26]

    Color Grading 101: Getting Started Color Grading for Editors, Cinematographers, Directors, and As- piring Colorists

    Charles Haine. Color Grading 101: Getting Started Color Grading for Editors, Cinematographers, Directors, and As- piring Colorists. Routledge, 2019. 2

  19. [27]

    The origins and prevalence of texture bias in convolutional neu- ral networks

    Katherine Hermann, Ting Chen, and Simon Kornblith. The origins and prevalence of texture bias in convolutional neu- ral networks. Advances in Neural Information Processing Systems, 33:19000–19015, 2020. 6

  20. [28]

    Deep preset: Blending and re- touching photos with color style transfer

    Man M Ho and Jinjia Zhou. Deep preset: Blending and re- touching photos with color style transfer. In Proceedings of the IEEE/CVF Winter Conference on Applications of Com- puter Vision, pages 2113–2121, 2021. 2, 5

  21. [29]

    Domain-aware universal style transfer

    Kibeom Hong, Seogkyu Jeon, Huan Yang, Jianlong Fu, and Hyeran Byun. Domain-aware universal style transfer. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 14609–14617, 2021. 2

  22. [30]

    Animate anyone: Consistent and controllable image- to-video synthesis for character animation

    Li Hu. Animate anyone: Consistent and controllable image- to-video synthesis for character animation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8153–8163, 2024. 4

  23. [31]

    Real- time neural style transfer for videos

    Haozhi Huang, Hao Wang, Wenhan Luo, Lin Ma, Wen- hao Jiang, Xiaolong Zhu, Zhifeng Li, and Wei Liu. Real- time neural style transfer for videos. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 783–791, 2017. 2

  24. [32]

    Arbitrary style transfer in real-time with adaptive instance normalization

    Xun Huang and Serge Belongie. Arbitrary style transfer in real-time with adaptive instance normalization. In Proceed- ings of the IEEE International Conference on Computer Vi- sion, pages 1501–1510, 2017. 2

  25. [33]

    Percep- tual losses for real-time style transfer and super-resolution

    Justin Johnson, Alexandre Alahi, and Li Fei-Fei. Percep- tual losses for real-time style transfer and super-resolution. In Computer Vision–ECCV 2016: 14th European Confer- ence, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part II 14, pages 694–711. Springer, 2016. 2

  26. [34]

    Imagic: Text-based real image editing with diffusion models

    Bahjat Kawar, Shiran Zada, Oran Lang, Omer Tov, Huiwen Chang, Tali Dekel, Inbar Mosseri, and Michal Irani. Imagic: Text-based real image editing with diffusion models. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023. 4

  27. [35]

    Neural preset for color style transfer

    Zhanghan Ke, Yuhao Liu, Lei Zhu, Nanxuan Zhao, and Ryn- son WH Lau. Neural preset for color style transfer. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14173–14182, 2023. 2

  28. [36]

    Color and Mastering for Digital Cinema

    Glenn Kennel. Color and Mastering for Digital Cinema . Routledge, 2006. 2

  29. [37]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 4015–4026, 2023. 8

  30. [38]

    Color space conversion for linear color grad- ing

    Dah-Jye Lee. Color space conversion for linear color grad- ing. In Intelligent Robots and Computer Vision XIX: Algo- rithms, Techniques, and Active Vision, pages 358–366. SPIE,

  31. [39]

    Combining markov random fields and convolutional neural networks for image synthesis

    Chuan Li and Michael Wand. Combining markov random fields and convolutional neural networks for image synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2479–2486, 2016. 2

  32. [40]

    Learning linear transformations for fast image and video style transfer

    Xueting Li, Sifei Liu, Jan Kautz, and Ming-Hsuan Yang. Learning linear transformations for fast image and video style transfer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 3809– 3817, 2019. 2

  33. [41]

    Diversified texture synthesis with feed-forward networks

    Yijun Li, Chen Fang, Jimei Yang, Zhaowen Wang, Xin Lu, and Ming-Hsuan Yang. Diversified texture synthesis with feed-forward networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 3920–3928, 2017. 2

  34. [42]

    A closed-form solution to photorealistic image stylization

    Yijun Li, Ming-Yu Liu, Xueting Li, Ming-Hsuan Yang, and Jan Kautz. A closed-form solution to photorealistic image stylization. In Proceedings of the European Conference on Computer Cision (ECCV), pages 453–468, 2018. 2

  35. [43]

    Adacm: adaptive colormlp for real-time universal photo-realistic style transfer

    Tianwei Lin, Honglin Lin, Fu Li, Dongliang He, Wenhao Wu, Meiling Wang, Xin Li, and Yong Liu. Adacm: adaptive colormlp for real-time universal photo-realistic style transfer. In Proceedings of the AAAI Conference on Artificial Intelli- gence, pages 1613–1621, 2023. 3, 6

  36. [44]

    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. 3

  37. [45]

    Decoupled weight decay regularization

    I Loshchilov. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 4

  38. [46]

    Deep video harmonization with color map- ping consistency

    Xinyuan Lu, Shengyuan Huang, Li Niu, Wenyan Cong, and Liqing Zhang. Deep video harmonization with color map- ping consistency. arXiv preprint arXiv:2205.00687 , 2022. 3

  39. [47]

    Deep photo style transfer

    Fujun Luan, Sylvain Paris, Eli Shechtman, and Kavita Bala. Deep photo style transfer. In Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition, pages 4990–4998, 2017. 2

  40. [48]

    Blind/referenceless image spatial quality evaluator

    Anish Mittal, Anush K Moorthy, and Alan C Bovik. Blind/referenceless image spatial quality evaluator. In 2011 conference record of the forty fifth asilomar conference on signals, systems and computers (ASILOMAR) , pages 723–

  41. [49]

    Edgeconnect: Generative image inpainting with adversarial edge learning

    Kamyar Nazeri, Eric Ng, Tony Joseph, Faisal Z Qureshi, and Mehran Ebrahimi. Edgeconnect: Generative image inpainting with adversarial edge learning. arXiv preprint arXiv:1901.00212, 2019. 8

  42. [50]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023. 7

  43. [51]

    Kinetic typography diffusion model

    Seonmi Park, Inhwan Bae, Seunghyun Shin, and Hae-Gon Jeon. Kinetic typography diffusion model. InEuropean Con- ference on Computer Vision, pages 166–185. Springer, 2024. 7

  44. [52]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems , 2019. 4

  45. [53]

    N- dimensional probability density function transfer and its ap- plication to color transfer

    Francois Pitie, Anil C Kokaram, and Rozenn Dahyot. N- dimensional probability density function transfer and its ap- plication to color transfer. In Tenth IEEE International Conference on Computer Vision (ICCV’05) Volume 1, pages 1434–1439. IEEE, 2005. 2

  46. [54]

    Au- tomated colour grading using colour distribution transfer

    Franc ¸ois Piti´e, Anil C Kokaram, and Rozenn Dahyot. Au- tomated colour grading using colour distribution transfer. Computer Vision and Image Understanding , 107(1-2):123– 137, 2007. 1, 2

  47. [55]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

  48. [56]

    Sam 2: Segment anything in images and videos

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman R¨adle, Chloe Rolland, Laura Gustafson, et al. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:2408.00714, 2024. 8

  49. [57]

    Color transfer between images

    Erik Reinhard, Michael Adhikhmin, Bruce Gooch, and Peter Shirley. Color transfer between images. IEEE Computer graphics and applications, 21(5):34–41, 2001. 2

  50. [58]

    Cognitive psychology for deep neural networks: A shape bias case study

    Samuel Ritter, David GT Barrett, Adam Santoro, and Matt M Botvinick. Cognitive psychology for deep neural networks: A shape bias case study. InInternational Conference on Ma- chine Learning, pages 2940–2949. PMLR, 2017. 6

  51. [59]

    Color grading techniques in different film genres

    Justin Robinson. Color grading techniques in different film genres. https://postprolist.com/article/color-grading- film-genres, 2024. 7

  52. [60]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10684–10695, 2022. 4, 6, 7

  53. [61]

    Artistic style transfer for videos and spherical images

    Manuel Ruder, Alexey Dosovitskiy, and Thomas Brox. Artistic style transfer for videos and spherical images. Inter- national Journal of Computer Vision , 126(11):1199–1219,

  54. [62]

    Parasite cinematography analysis (in- depth)

    Salik Waquas Shaw. Parasite cinematography analysis (in- depth). https://colorculture.org/parasite- cinematography-analysis/, 2024. 1

  55. [63]

    Task- specific scene structure representations

    Jisu Shin, Seunghyun Shin, and Hae-Gon Jeon. Task- specific scene structure representations. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 2272– 2281, 2023. 8

  56. [64]

    Close imitation of expert retouch- ing for black-and-white photography

    Seunghyun Shin, Jisu Shin, Jihwan Bae, Inwook Shim, and Hae-Gon Jeon. Close imitation of expert retouch- ing for black-and-white photography. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 25037–25046, 2024. 2, 7

  57. [65]

    Neighbor relations matter in video scene detection

    Jiawei Tan, Hongxing Wang, Jiaxin Li, Zhilong Ou, and Zhangbin Qian. Neighbor relations matter in video scene detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 18473– 18482, 2024. 8

  58. [66]

    Neural discrete representation learning

    Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. Advances in Neural Information Processing Systems, 30, 2017. 4

  59. [67]

    Exploratory study and analysis on color grading technique for horror short movie

    Rayden Vebrianto and Jimmy Pratama. Exploratory study and analysis on color grading technique for horror short movie. IJISTECH International Journal of Information Sys- tem and Technology(IJISTECH), 6(5):603–612, 2023. 7

  60. [68]

    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, pages 2471–2480, 2021. 3

  61. [69]

    Consistent video style transfer via relaxation and regular- ization

    Wenjing Wang, Shuai Yang, Jizheng Xu, and Jiaying Liu. Consistent video style transfer via relaxation and regular- ization. IEEE Transactions on Image Processing, 29:9125– 9139, 2020. 2

  62. [70]

    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 Process- ing, 13(4):600–612, 2004. 5

  63. [71]

    Transferring color to greyscale images

    Tomihisa Welsh, Michael Ashikhmin, and Klaus Mueller. Transferring color to greyscale images. In Proceedings of the 29th annual conference on Computer graphics and inter- active techniques, pages 277–280, 2002. 2

  64. [72]

    Cap- vstnet: content affinity preserved versatile style transfer

    Linfeng Wen, Chengying Gao, and Changqing Zou. Cap- vstnet: content affinity preserved versatile style transfer. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, 2023. 2, 5

  65. [73]

    Scene consis- tency representation learning for video scene segmentation

    Haoqian Wu, Keyu Chen, Yanan Luo, Ruizhi Qiao, Bo Ren, Haozhe Liu, Weicheng Xie, and Linlin Shen. Scene consis- tency representation learning for video scene segmentation. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 14021–14030...

  66. [74]

    Preserving global and local temporal consistency for arbitrary video style transfer

    Xinxiao Wu and Jialu Chen. Preserving global and local temporal consistency for arbitrary video style transfer. In Proceedings of the 28th ACM International Conference on Multimedia, pages 1791–1799, 2020. 2

  67. [75]

    Ccpl: Con- trastive coherence preserving loss for versatile style transfer

    Zijie Wu, Zhen Zhu, Junping Du, and Xiang Bai. Ccpl: Con- trastive coherence preserving loss for versatile style transfer. In European Conference on Computer Vision , pages 189–

  68. [76]

    Joint bilateral learning for real-time universal photorealistic style transfer

    Xide Xia, Meng Zhang, Tianfan Xue, Zheng Sun, Hui Fang, Brian Kulis, and Jiawen Chen. Joint bilateral learning for real-time universal photorealistic style transfer. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision, pages 327–342. Springer, 2020. 2

  69. [77]

    Real-time localized photorealistic video style transfer

    Xide Xia, Tianfan Xue, Wei-sheng Lai, Zheng Sun, Abby Chang, Brian Kulis, and Jiawen Chen. Real-time localized photorealistic video style transfer. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 1089–1098, 2021. 2

  70. [78]

    Aggregated residual transformations for deep neural networks

    Saining Xie, Ross Girshick, Piotr Doll ´ar, Zhuowen Tu, and Kaiming He. Aggregated residual transformations for deep neural networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1492–1500,

  71. [79]

    Frame difference-based temporal loss for video stylization

    Jianjin Xu, Zheyang Xiong, and Xiaolin Hu. Frame difference-based temporal loss for video stylization. arXiv preprint arXiv:2102.05822, 2021. 5

  72. [80]

    Magicanimate: Temporally consistent human im- age animation using diffusion model

    Zhongcong Xu, Jianfeng Zhang, Jun Hao Liew, Hanshu Yan, Jia-Wei Liu, Chenxu Zhang, Jiashi Feng, and Mike Zheng Shou. Magicanimate: Temporally consistent human im- age animation using diffusion model. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Reco...

  73. [81]

    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, pages 17522–17531, 2022. 3

  74. [82]

    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 European Con- ference on Computer Vision, pages 201–217. Springer, 2022. 3

  75. [83]

    Photorealistic style transfer via wavelet transforms

    Jaejun Yoo, Youngjung Uh, Sanghyuk Chun, Byeongkyu Kang, and Jung-Woo Ha. Photorealistic style transfer via wavelet transforms. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 9036–9045,

  76. [84]

    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 , 44(4):2058– 2073, 2020. 3

  77. [85]

    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. 3

  78. [86]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3836–3847, 2023. 6

  79. [87]

    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/CVF Conference on Computer Vision and Pattern Recognition, pages 586–595, 2018. 5

  80. [206]

    Springer, 2022. 2, 5

Pith tools

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