pith. sign in

arxiv: 2606.01819 · v1 · pith:KBIUZMBRnew · submitted 2026-06-01 · 💻 cs.CV · eess.IV

Hist2Style: Histogram-Guided Stylization with Bilateral Grids

Dekel Galor , Adam Pikielny , Zhoutong Zhang , Ke Wang , Laura Waller , Jiawen Chen , Ilya Chugunov This is my paper

Pith reviewed 2026-06-28 15:21 UTC · model grok-4.3

classification 💻 cs.CV eess.IV
keywords photorealistic style transferbilateral gridshistogram guidanceimage stylizationreal-time editingcolor transferedge-aware transformslightweight networks
0
0 comments X

The pith

Hist2Style distills large editing models into a lightweight bilateral-grid network conditioned on style histograms to deliver real-time photorealistic stylization without hallucinations.

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

The paper presents Hist2Style as a way to achieve photorealistic style transfer that matches an input image's color and tone to a target while keeping original scene details. Large generative models can do similar edits but run slowly, risk hallucinations, and offer little user control, making them impractical for high-resolution or interactive work. Hist2Style instead trains a compact network on data from language and vision-language models to learn locally affine transforms inside bilateral space. Conditioning the network on a histogram embedding of the style target gives an interpretable way to tweak output colors and tones directly. This construction keeps content structure intact by design and enables real-time, high-resolution results with user adjustments.

Core claim

Hist2Style formulates stylization as locally affine transforms performed in bilateral space and conditions those transforms on a histogram-based embedding of the style target. Training a lightweight network on a large supervised corpus generated by language and vision-language models produces a model that executes edge-aware color and tone edits while preserving visual fidelity.

What carries the argument

The bilateral-grid formulation that applies locally affine transforms in bilateral space, conditioned on a histogram embedding of the style target.

If this is right

  • Real-time processing becomes feasible for high-resolution images.
  • Users can interactively adjust output color and tone by editing the target histogram.
  • Content structure is preserved by the bilateral-grid constraints without additional regularization.
  • Hallucinations are avoided because operations stay within locally affine color transforms.

Where Pith is reading between the lines

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

  • The histogram interface could be adapted to control other spatially varying edits such as local contrast or exposure.
  • Similar distillation into bilateral grids might speed up related tasks like video color grading if temporal consistency is added.
  • The approach suggests that many global appearance edits can be reduced to histogram-guided local affine adjustments without full generative models.

Load-bearing premise

The locally affine bilateral-grid transforms distilled from the supervised corpus will generalize to arbitrary real-world images without introducing artifacts or fidelity loss.

What would settle it

Applying the trained network to real images outside the generated training distribution and observing either changes to scene structure or visible artifacts would falsify the generalization claim.

Figures

Figures reproduced from arXiv: 2606.01819 by Adam Pikielny, Dekel Galor, Ilya Chugunov, Jiawen Chen, Ke Wang, Laura Waller, Zhoutong Zhang.

Figure 1
Figure 1. Figure 1: Our method preserves content and detail by mapping [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Selective distillation to condense the capabilities of a large image editing model into a lightweight, specialized network for photorealistic stylization. We procedurally generate style editing prompts with a large language model. The prompts are used to instruct an editing model to edit standard images into different style variations, which are coherent across images. Our model is then trained to mimic th… view at source ↗
Figure 3
Figure 3. Figure 3: Model architecture (left). Our model separately embeds a downsampled content image and style color histogram via ConvNeXt blocks. The two are then fused with cross attention, and fed through the output head to produce a spatially adaptive color transform known as the bilateral grid. The bilateral grid is applied to the content image and fed through a learned per-channel nonlinearity (LUT) to produce a styl… view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative comparisons with baseline methods on images from the user-study evaluation set (Tab. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative comparisons with baseline methods on an independently collected, non-public dataset, highlighting Hist2Style’s [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: User control. We can interactively control the stylization process by directly editing the guiding histogram. This global user intent is fed to Hist2Style, which applies local changes that are adaptive to the image being edited. Users can edit the histograms directly by dragging the curve or use more familiar operations via custom sliders (see Sec. 3.6 for more details). During training, for each content i… view at source ↗
Figure 7
Figure 7. Figure 7: Stylization Quality Assessment (SQA) metric. We in￾troduce SQA, a new metric for assessing stylization quality which is aligned to human preference, as shown in the correlation plot. For clarity, metrics were oriented so that higher is better [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Ablation studies and extensions. We explored differ￾ent architectural and training configurations. ”Color Space Loss” refers to the output from a model where the loss was applied to raw color space during training, instead of perceptual space. Compared to this model, Hist2Style expresses a richer color transfer and bet￾ter maintains image contrast. ”Robust” refers to the output from a model trained not wit… view at source ↗
read the original abstract

Photorealistic style transfer aims to match the color and tone of an input image to that of a style target while preserving the content and details of the original scene. Although existing large image models can facilitate these kinds of appearance edits, their high computational demands, potential for hallucinations, and limited user control make them unsuitable for high-resolution, real-time workflows. We introduce Hist2Style, a bilateral-grid formulation for fast, edge-aware stylization that preserves visual fidelity by constraining operations to locally affine transforms in bilateral space. Our model distills a large image editing model into a lightweight network by training on a large supervised corpus generated with language and vision-language models, targeting spatially varying color edits. The network conditions on a histogram-based embedding of the style target to provide an interpretable interface for adjusting the output style by modifying the target color distribution. Overall, Hist2Style maintains content structure by construction, avoids hallucinations, and supports real-time, high-resolution photorealistic stylization with interactive user-controllable color and tone adjustments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript introduces Hist2Style, a bilateral-grid formulation for photorealistic style transfer. It distills a large image editing model into a lightweight network trained on a supervised corpus generated by language and vision-language models, targeting spatially varying color edits. The network conditions on a histogram-based embedding of the style target to enable interpretable user adjustments to color and tone. The approach claims to maintain content structure by construction via locally affine transforms in bilateral space, avoid hallucinations, and support real-time, high-resolution stylization with interactive control.

Significance. If the claims hold, the work would be significant for enabling efficient, controllable photorealistic stylization without the computational cost or hallucination risks of large models. The bilateral-grid approach for edge-aware, locally affine transforms is a strength for real-time high-resolution workflows, and the histogram conditioning provides a practical interface for user control. The distillation strategy from large models to a lightweight network could have broader applicability.

major comments (2)
  1. [Abstract] Abstract: The claim that content structure is 'maintained by construction' and hallucinations are 'avoided' rests on the locally affine transforms in bilateral space, but the abstract provides no equations, derivation, or analysis showing why these transforms remain artifact-free when input histograms and content deviate from the synthetic training corpus generated by LMs/VLMs. This is load-bearing for the central no-hallucination and generalization claims.
  2. [Method] The skeptic concern on generalization is valid here: nothing in the formulation as described guarantees that the predicted per-grid affine coefficients will not over-smooth or introduce ringing at boundaries unseen during distillation. The paper must provide either a theoretical argument or targeted experiments demonstrating robustness outside the training distribution.
minor comments (1)
  1. [Abstract] Abstract: The phrase 'targeting spatially varying color edits' is vague; specify how the histogram embedding enforces spatial variation versus global tone shifts.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major comment below and commit to appropriate revisions.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that content structure is 'maintained by construction' and hallucinations are 'avoided' rests on the locally affine transforms in bilateral space, but the abstract provides no equations, derivation, or analysis showing why these transforms remain artifact-free when input histograms and content deviate from the synthetic training corpus generated by LMs/VLMs. This is load-bearing for the central no-hallucination and generalization claims.

    Authors: We agree the abstract would benefit from added context. The structure-preserving property follows from the bilateral grid being constructed directly from the content image's spatial and intensity coordinates; locally affine transforms are then applied within these content-adaptive cells, which by definition prevents cross-edge mixing or structural hallucination regardless of the specific coefficient values (the operation remains a per-region color/tone adjustment). We will revise the abstract to briefly note this property and reference the equations in Section 3. revision: yes

  2. Referee: [Method] The skeptic concern on generalization is valid here: nothing in the formulation as described guarantees that the predicted per-grid affine coefficients will not over-smooth or introduce ringing at boundaries unseen during distillation. The paper must provide either a theoretical argument or targeted experiments demonstrating robustness outside the training distribution.

    Authors: The concern is valid; while the bilateral formulation limits certain artifacts (e.g., no edge crossing), network-predicted coefficients could still produce over-smoothing or ringing on out-of-distribution inputs. We will add a short theoretical note on the bounded interpolation properties of the bilateral grid and include targeted experiments on challenging OOD cases (extreme histograms, unseen fine structures) to demonstrate robustness. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The provided abstract and description contain no equations, derivations, or load-bearing steps that reduce to inputs by construction. The claim of preserving structure 'by construction' via locally affine bilateral-grid transforms is a direct statement of the method's design property rather than a self-referential reduction. Distillation from external LMs/VLMs is described as a training process without evidence of fitted parameters being renamed as predictions or self-citation chains. No self-definitional, fitted-input, or uniqueness-imported patterns are identifiable from the text.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that locally affine transforms in bilateral space suffice to capture spatially varying color edits without content alteration; no free parameters or invented entities are explicitly named in the abstract.

axioms (1)
  • domain assumption Locally affine transforms in bilateral space maintain content structure by construction
    Directly stated in the abstract as the mechanism that preserves visual fidelity and avoids hallucinations.

pith-pipeline@v0.9.1-grok · 5727 in / 1164 out tokens · 30406 ms · 2026-06-28T15:21:51.850763+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

60 extracted references · 10 linked inside Pith

  1. [1]

    Accessed: 2025- 11-13

    Unsplash.https://unsplash.com. Accessed: 2025- 11-13

    2025

  2. [2]

    Gradio: Hassle-free sharing and testing of ml models in the wild.arXiv preprint arXiv:1906.02569, 2019

    Abubakar Abid, Ali Abdalla, Ali Abid, Dawood Khan, Ab- dulrahman Alfozan, and James Zou. Gradio: Hassle-free sharing and testing of ml models in the wild.arXiv preprint arXiv:1906.02569, 2019

    Pith/arXiv arXiv 1906

  3. [3]

    Fast high- dimensional filtering using the permutohedral lattice.Euro- graphics, 2010

    Andrew Adams, Jongmin Baek, and Abe Davis. Fast high- dimensional filtering using the permutohedral lattice.Euro- graphics, 2010

    2010

  4. [4]

    Adobe photoshop lightroom, 2007

    Adobe Systems Incorporated. Adobe photoshop lightroom, 2007

    2007

  5. [5]

    The fast bilateral solver

    Jonathan T Barron and Ben Poole. The fast bilateral solver. ECCV, 2016

    2016

  6. [6]

    Fast bilateral-space stereo for synthetic de- focus.CVPR, 2015

    Jonathan T Barron, Andrew Adams, YiChang Shih, and Car- los Hern´andez. Fast bilateral-space stereo for synthetic de- focus.CVPR, 2015

    2015

  7. [7]

    Experiment tracking with weights and bi- ases, 2020

    Lukas Biewald. Experiment tracking with weights and bi- ases, 2020

    2020

  8. [8]

    Tim Brooks, Aleksander Holynski, and Alexei A. Efros. In- structpix2pix: Learning to follow image editing instructions. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    2023

  9. [9]

    IQA-PyTorch: Pytorch toolbox for image quality assessment

    Chaofeng Chen and Jiadi Mo. IQA-PyTorch: Pytorch toolbox for image quality assessment. [Online]. Avail- able:https : / / github . com / chaofengc / IQA - PyTorch, 2022

    2022

  10. [10]

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

    2023

  11. [11]

    Photowct 2: Compact autoencoder for photorealistic style transfer resulting from blockwise training and skip connections of high-frequency residuals, 2021

    Tai-Yin Chiu and Danna Gurari. Photowct 2: Compact autoencoder for photorealistic style transfer resulting from blockwise training and skip connections of high-frequency residuals, 2021

    2021

  12. [12]

    Color psychology: Effects of perceiving color on psychological functioning in humans

    Andrew Elliot and Markus Maier. Color psychology: Effects of perceiving color on psychological functioning in humans. Annual review of psychology, 65, 2013

    2013

  13. [13]

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

    Afifi et al. Histogan: Controlling colors of gan-generated and real images via color histograms. InCVPR, 2021

    2021

  14. [14]

    Toward generalized image quality assessment: Relaxing the perfect reference quality assumption, 2025

    Chen et al. Toward generalized image quality assessment: Relaxing the perfect reference quality assumption, 2025

    2025

  15. [15]

    Image quality assessment: From human to machine preference, 2025

    Li et al. Image quality assessment: From human to machine preference, 2025

    2025

  16. [16]

    PyTorch Lightning, 2019

    William Falcon and The PyTorch Lightning team. PyTorch Lightning, 2019

    2019

  17. [17]

    Gatys, Alexander S

    Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge. A neural algorithm of artistic style.CoRR, abs/1508.06576, 2015

    Pith/arXiv arXiv 2015

  18. [18]

    Barron, Samuel W

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

    2017

  19. [19]

    Sa-lut: Spatial adaptive 4d look-up ta- ble for photorealistic style transfer, 2025

    Zerui Gong, Zhonghua Wu, Qingyi Tao, Qinyue Li, and Chen Change Loy. Sa-lut: Spatial adaptive 4d look-up ta- ble for photorealistic style transfer, 2025

    2025

  20. [20]

    Routledge, 2019

    Charles Haine.Color Grading 101: getting started color grading for editors, cinematographers, directors, and aspir- ing colorists. Routledge, 2019

    2019

  21. [21]

    Prompt-to-prompt im- age editing with cross attention control.arXiv preprint arXiv:2208.01626, 2022

    Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Prompt-to-prompt im- age editing with cross attention control.arXiv preprint arXiv:2208.01626, 2022

    Pith/arXiv arXiv 2022

  22. [22]

    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. InICCV, 2017

    2017

  23. [23]

    Matplotlib: A 2d graphics environment

    John D Hunter. Matplotlib: A 2d graphics environment. Computing in science & engineering, 9(03):90–95, 2007

    2007

  24. [24]

    Neural style transfer: A review.CoRR, abs/1705.04058, 2017

    Yongcheng Jing, Yezhou Yang, Zunlei Feng, Jingwen Ye, and Mingli Song. Neural style transfer: A review.CoRR, abs/1705.04058, 2017

    Pith/arXiv arXiv 2017

  25. [25]

    Zhanghan Ke, Yuhao Liu, Lei Zhu, Nanxuan Zhao, and Ryn- son W.H. Lau. Neural preset for color style transfer. InCom- puter Vision and Pattern Recognition Conference (CVPR), 2023

    2023

  26. [26]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization.arXiv preprint arXiv:1412.6980, 2014

    Pith/arXiv arXiv 2014

  27. [27]

    Flux.1 kontext: Flow matching for in-context image generation and editing in latent space, 2025

    Black Forest Labs, Stephen Batifol, Andreas Blattmann, Frederic Boesel, Saksham Consul, Cyril Diagne, Tim Dock- horn, Jack English, Zion English, Patrick Esser, Sumith Ku- lal, Kyle Lacey, Yam Levi, Cheng Li, Dominik Lorenz, Jonas M¨uller, Dustin Podell, Robin Rombach, Harry Saini, Axel Sauer, and Luke Smith. Flux.1 kontext: Flow matching for in-context i...

    2025

  28. [28]

    D-LUT: Photorealistic Style Transfer via Diffusion Process

    Mujing Li, Guanjie Wang, Xingguang Zhang, Qifeng Liao, and Chenxi Xiao. D-LUT: Photorealistic Style Transfer via Diffusion Process . In2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 9206–9214, Los Alamitos, CA, USA, 2025. IEEE Computer Society

    2025

  29. [29]

    Universal style transfer via feature transforms.CoRR, abs/1705.08086, 2017

    Yijun Li, Chen Fang, Jimei Yang, Zhaowen Wang, Xin Lu, and Ming-Hsuan Yang. Universal style transfer via feature transforms.CoRR, abs/1705.08086, 2017

    Pith/arXiv arXiv 2017

  30. [30]

    A closed-form solution to photorealistic image stylization.CoRR, abs/1802.06474, 2018

    Yijun Li, Ming-Yu Liu, Xueting Li, Ming-Hsuan Yang, and Jan Kautz. A closed-form solution to photorealistic image stylization.CoRR, abs/1802.06474, 2018

    Pith/arXiv arXiv 2018

  31. [31]

    A convnet for the 2020s.CoRR, abs/2201.03545, 2022

    Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feicht- enhofer, Trevor Darrell, and Saining Xie. A convnet for the 2020s.CoRR, abs/2201.03545, 2022. 9

    arXiv 2022

  32. [32]

    Deep photo style transfer.arXiv preprint arXiv:1703.07511, 2017

    Fujun Luan, Sylvain Paris, Eli Shechtman, and Kavita Bala. Deep photo style transfer.arXiv preprint arXiv:1703.07511, 2017

    Pith/arXiv arXiv 2017

  33. [33]

    Null-text inversion for editing real images using guided diffusion models, 2022

    Ron Mokady, Amir Hertz, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Null-text inversion for editing real images using guided diffusion models, 2022

    2022

  34. [34]

    pandas-dev/pandas: Pandas

    The pandas development team. pandas-dev/pandas: Pandas. 2020

    2020

  35. [35]

    On aliased resizing and surprising subtleties in gan evaluation

    Gaurav Parmar, Richard Zhang, and Jun-Yan Zhu. On aliased resizing and surprising subtleties in gan evaluation. InCVPR, 2022

    2022

  36. [36]

    Pytorch: An im- perative style, high-performance deep learning library.Ad- vances in Neural Information Processing Systems, 32, 2019

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An im- perative style, high-performance deep learning library.Ad- vances in Neural Information Processing Systems, 32, 2019

    2019

  37. [37]

    Scikit-learn: Machine learning in python.the Journal of machine Learning research, 12:2825–2830, 2011

    Fabian Pedregosa, Ga ¨el Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit-learn: Machine learning in python.the Journal of machine Learning research, 12:2825–2830, 2011

    2011

  38. [38]

    Computational optimal transport, 2020

    Gabriel Peyr ´e and Marco Cuturi. Computational optimal transport, 2020

    2020

  39. [39]

    P. F. Piti ´e and A. C. Kokaram. The linear monge- kantorovitch linear colour mapping for example-based colour transfer. InProceedings of the 4th IEE European Conference on Visual Media Production (CVMP), London, United Kingdom, 2007

    2007

  40. [40]

    P. F. Piti ´e, A. C. Kokaram, and R. Dahyot. N-dimensional probability density function transfer and its application to colour transfer. InProceedings of the IEEE International Conference on Computer Vision (ICCV), Beijing, China, 2005

    2005

  41. [41]

    P. F. Piti ´e, A. C. Kokaram, and R. Dahyot. Towards auto- mated colour grading. InProceedings of the 2nd IEE Euro- pean Conference on Visual Media Production (CVMP), Lon- don, United Kingdom, 2005

    2005

  42. [42]

    P. F. Piti´e, A. C. Kokaram, and R. Dahyot. Automated colour grading using colour distribution transfer.Computer Vision and Image Understanding, 2007

    2007

  43. [43]

    P. F. Piti ´e, A. C. Kokaram, and R. Dahyot. Enhancement of digital photographs using color transfer techniques. In Single-Sensor Imaging, pages 295–321. CRC Press, 2008

    2008

  44. [44]

    Julien Porquet, Sitong Wang, and Lydia B. Chilton. Copying style, extracting value: Illustrators’ perception of AI style transfer and its impact on creative labor, 2025

    2025

  45. [45]

    Reinhard, M

    E. Reinhard, M. Ashikhmin, B. Gooch, and P. Shirley. Color transfer between images.IEEE Computer Graphics and Ap- plications, 21(5):34–41, 2001

    2001

  46. [46]

    statsmodels: Econo- metric and statistical modeling with python

    Skipper Seabold and Josef Perktold. statsmodels: Econo- metric and statistical modeling with python. In9th Python in Science Conference, 2010

    2010

  47. [47]

    A survey of multimodal-guided image editing with text-to-image diffu- sion models, 2024

    Xincheng Shuai, Henghui Ding, Xingjun Ma, Rongcheng Tu, Yu-Gang Jiang, and Dacheng Tao. A survey of multimodal-guided image editing with text-to-image diffu- sion models, 2024

    2024

  48. [48]

    Very deep convo- lutional networks for large-scale image recognition.arXiv preprint arXiv:1409.1556, 2014

    Karen Simonyan and Andrew Zisserman. Very deep convo- lutional networks for large-scale image recognition.arXiv preprint arXiv:1409.1556, 2014

    Pith/arXiv arXiv 2014

  49. [49]

    Gem- ini: A family of highly capable multimodal models, 2025

    Gemini Team, Rohan Anil, Sebastian Borgeaud, et al. Gem- ini: A family of highly capable multimodal models, 2025

    2025

  50. [50]

    Unsplash lite dataset photos

    Unsplash. Unsplash lite dataset photos. GitHub repository

  51. [51]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is all you need.CoRR, abs/1706.03762, 2017

    Pith/arXiv arXiv 2017

  52. [52]

    Image editing with diffusion models: A sur- vey, 2025

    Jia Wang, Jie Hu, Xiaoqi Ma, Hanghang Ma, Xiaoming Wei, and Enhua Wu. Image editing with diffusion models: A sur- vey, 2025

    2025

  53. [53]

    Bilateral guided radiance field processing, 2024

    Yuehao Wang, Chaoyi Wang, Bingchen Gong, and Tianfan Xue. Bilateral guided radiance field processing, 2024

    2024

  54. [54]

    Seaborn: statistical data visualization

    Michael L Waskom. Seaborn: statistical data visualization. Journal of Open Source Software, 6(60):3021, 2021

    2021

  55. [55]

    Dimitrova, and Shih-Fu Chang

    Cheng-Yu Wei, N. Dimitrova, and Shih-Fu Chang. Color- mood analysis of films based on syntactic and psychological models. In2004 IEEE International Conference on Multi- media and Expo (ICME) (IEEE Cat. No.04TH8763), pages 831–834 V ol.2, 2004

    2004

  56. [56]

    Qwen-image technical report, 2025

    Chenfei Wu, Jiahao Li, Jingren Zhou, Junyang Lin, Kaiyuan Gao, Kun Yan, Sheng ming Yin, Shuai Bai, Xiao Xu, Yilei Chen, Yuxiang Chen, Zecheng Tang, Zekai Zhang, Zhengyi Wang, An Yang, Bowen Yu, Chen Cheng, Dayiheng Liu, De- qing Li, Hang Zhang, Hao Meng, Hu Wei, Jingyuan Ni, Kai Chen, Kuan Cao, Liang Peng, Lin Qu, Minggang Wu, Peng Wang, Shuting Yu, Tingk...

    2025

  57. [57]

    Joint bilateral learning for real-time universal photorealistic style transfer.CoRR, abs/2004.10955, 2020

    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.CoRR, abs/2004.10955, 2020

    arXiv 2004

  58. [58]

    Photorealistic style transfer via wavelet transforms

    Jaejun Yoo, Youngjung Uh, Sanghyuk Chun, Byeongkyu Kang, and Jung-Woo Ha. Photorealistic style transfer via wavelet transforms. InInternational Conference on Com- puter Vision (ICCV), 2019

    2019

  59. [59]

    Style transfer: A decade survey, 2025

    Tianshan Zhang and Hao Tang. Style transfer: A decade survey, 2025

    2025

  60. [60]

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks.CoRR, abs/1703.10593, 2017. 10

    arXiv 2017