Pith. sign in

REVIEW 3 major objections 5 minor 61 references

One-shot Face Sketch Synthesis in the Wild via Generative Diffusion Prior and Instruction Tuning

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read A single photo-sketch pair is enough for a frozen diffusion model to synthesize realistic face sketches in the wild.

desk verdict OS-Sketch is a useful benchmark, but the point estimates in Table 1 don't yet carry the paper's one-shot superiority claim. read the letter →

arxiv 2506.15312 v1 pith:O2KHKJI4 submitted 2025-06-18 cs.GR cs.CRcs.CVcs.CY

classification cs.GRcs.CRcs.CVcs.CY
keywords facesketchsynthesisone-shotlearningdiffusionmodelstextinstructionoptimizationtuningimage-to-imagetranslationout-of-distributiongeneralizationphoto-sketchbenchmark
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that face sketch synthesis can be reduced to a one-shot learning problem: with a single aligned photo-sketch pair, a frozen text-to-image diffusion model can be steered into a face-sketch generator by optimizing a text instruction. The instruction is found by gradient descent on two losses, one that teaches the model to reconstruct the sketch from the photo and one that aligns the instruction with the photo-to-sketch direction in an image-text embedding space. The authors build OS-Sketch, a benchmark of 400 photo-sketch pairs with varied sketch styles and in-the-wild photos, and they select one pair for training and the rest for testing. On that benchmark the method is reported to produce realistic, identity-consistent sketches and to outperform previous face-sketch and text-driven diffusion baselines on SSIM, LPIPS, and FID. The value of the claim, if correct, is that hand-drawn sketch data no longer has to be collected at scale.

What carries the argument

The load-bearing object is the learned text instruction $C_{\mathrm{Text}}$, a short token sequence optimized by gradient descent while the denoising network stays frozen. It is trained with two losses: a reconstruction loss ($L_{\mathrm{rec}}$, Eq. 3) that measures how well the denoiser can recover the sketch latent from the noisy target conditioned on the encoded photo, and an editing-direction loss ($L_{\mathrm{edit}}$, Eq. 5) that maximizes cosine similarity between the instruction embedding and the embedding-space displacement from the photo to the sketch. A hybrid instruction formed by the optimized part $\langle ins \rangle$ plus a manually supplied $\langle optional \rangle$ part controls tone and background at inference. The mechanism matters because it lets one visual example stand for the whole photo-to-sketch mapping, transferring the task description into a form the frozen model already understands.

What would settle it

Run the model on a set of face photos that are deliberately extreme in pose, illumination, or occlusion; if any of them fail to preserve facial identity or produce sketch-like textures while a standard multi-pair baseline succeeds, the one-shot generalization claim is refuted.

Watch

Extended reading notes

Core claim

In the paper's own terms, the central discovery is that the editing direction from a face photo to a sketch can be captured in a single text embedding. Given one pair $\{I_{\mathrm{photo}}, I_{\mathrm{sketch}}\}$, the method optimizes the text instruction $C_{\mathrm{Text}}$ so that denoising the noisy sketch latent conditioned on the encoded photo reconstructs the sketch, while the instruction's embedding also moves in the photo-to-sketch direction measured by cosine similarity. The diffusion model itself is never fine-tuned; only the instruction changes. At inference the optimized instruction can be concatenated with optional textual hints, and the frozen model converts new photos, including out-of-distribution internet photos, into sketches that inherit the style and tone of the single training sketch. The paper reports best SSIM, LPIPS, and FID on OS-Sketch compared with prior one-shot-trained baselines.

Load-bearing premise

The entire method rests on the assumption that a frozen text-to-image diffusion model already knows enough about faces and sketching that one carefully chosen text instruction can make it produce a good sketch from any new face photo.

Editorial extensions

If this is right

  • Artists or users can adapt a sketch generator to a new style by providing exactly one example, so dataset collection and manual labeling shrink to a single pair.
  • The same instruction-optimization recipe may transfer to other image-to-image style tasks that a text-to-image diffusion model can already approximate, not only face sketches.
  • Because the diffusion model is frozen, per-task training cost is bounded by instruction optimization rather than full model fine-tuning.
  • Style control becomes a text-level operation: the learned instruction can be supplemented with optional textual hints to set color, background, or tone.
  • On the new OS-Sketch benchmark, the authors' method beats conventional face-sketch methods that collapse in the one-shot setting, suggesting that one-shot evaluation is a viable replacement for large paired training.

Reading between the lines

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

  • An implicit constraint is that the diffusion prior must already contain the face-to-sketch concept; the method is a retrieval of that prior via text, not a new representation of sketching. If a sketch style is far outside the prior, one optimized pair will likely not recover it.
  • A natural test is to measure per-sample quality on the 40 internet-collected out-of-distribution pairs separately; if performance there is much worse than on in-distribution pairs, the in-the-wild generalization claim would be weakened.
  • The same optimization could be applied with more than one training pair for styles that fail one-shot; whether performance scales smoothly with additional pairs is an open question the paper does not address.
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

3 major / 5 minor

Summary. The paper presents a one-shot face sketch synthesis method based on a frozen text-to-image diffusion model. Given a single photo-sketch pair, the method optimizes a text instruction embedding C_Text by minimizing a diffusion reconstruction loss (Eq. 3) on the sketch conditioned on the photo latent plus a CLIP edit-direction loss (Eq. 5) between the learned instruction and the photo-to-sketch CLIP embedding difference. At inference, the learned instruction can be concatenated with an optional manually supplied text prompt (Eq. 7). The authors introduce OS-Sketch, a 400-pair benchmark with multiple sketch styles and in-the-wild photos, and report quantitative comparisons (SSIM, LPIPS, FID) and qualitative results against face-sketch and diffusion-editing baselines.

Significance. If the reported results are reproducible across training-pair choices, the method would be a practically useful contribution: it removes the need for large paired training sets, enables style transfer from a single exemplar, and makes use of a strong pretrained diffusion prior. The paper also contributes a new benchmark and includes ablations of the edit-direction loss and hybrid-instruction component. The evaluation is external to the training pair and the optimization objective is clearly specified. However, the central empirical claim of significant superiority is not yet backed by the evidence as presented: Table 1 is a point estimate from one unspecified training pair, the extra inference prompt is not controlled across methods, and the dependence on the frozen prior's capacity is acknowledged in the paper's own limitation section but not analyzed. These are fixable with additional experiments.

major comments (3)
  1. [Section 4.2, Table 1] The claim that the method 'significantly improves upon previous methods' rests on point estimates SSIM 0.461, LPIPS 0.348, and FID 68.46, but the paper never states which photo-sketch pair was used for the one-shot training, how many optimization runs were averaged, or how the choice of exemplar affects the metrics. The OOD experiment in Section 4.2 ('we select 4 original sketches ... for training and testing the rest') confirms that the exemplar is a free choice, and in one-shot learning the exemplar sets both the sketch style and the photo domain. Because FID over 399 generated versus 399 real images has substantial sampling noise, the gaps over HIDA (79.13) and InstructPix2Pix (78.91) may not be significant. Please report mean plus/minus standard deviation across several training pairs and runs, state the training-pair selection protocol, and give the number of samples used in the FID computation.
  2. [Section 4.2 and Eq. (7)] The quantitative comparison uses an extra inference prompt, 'covert the image color to black and white with a white background,' but it is not stated whether this prompt was given to the text-driven baselines or only to the proposed method through the <optional> component. Since the ablation in Fig. 9 shows that removing this prompt causes 'deviations' in tone, the prompt is load-bearing for the reported quality, and a comparison that gives it only to one method would not isolate the learned instruction. Please clarify the protocol and report results both with and without the optional text for every text-conditioned method.
  3. [Section 3.2, Eq. (3), and Section 5] The method optimizes only C_Text while the denoising network epsilon_theta is frozen, so its success depends on the pretrained text-to-image model already being able to represent fine sketch textures and the photo-to-sketch mapping. The paper's own Limitation section concedes that 'there may be limitations in handling all types of sketch styles or achieve more precise sketch textures due to the operational mode of text-to-image diffusion models,' and Table 1 aggregates over all 400 pairs without a per-style breakdown. Please add per-cluster quantitative results (e.g., treating the 40 internet sketches separately) and a failure-case analysis; without this, the claimed generality across sketch styles is not evidenced.
minor comments (5)
  1. [Section 4.1] The pretrained model used is described only as 'a frozen pre-trained model'; please specify the exact stable-diffusion checkpoint, CLIP model, and image resolution, because the method's capacity claim depends on these choices.
  2. [Sections 4.2 and 4.3] The word 'covert' in 'covert the image color to black and white' should be 'convert'; the same typo appears in Section 4.3.
  3. [Section 4.2] FID computation details are missing; specify the number of generated and real images, whether standard Inception-v3 features are used, and the resolution at which outputs are compared.
  4. [Fig. 3 caption] The set notation includes a stray comma in '{I_photo, I_sketch,}', and the three panels are described in an order that does not match the panel labels; please correct.
  5. [Section 3.1, Eqs. (1) and (3)] The instruction variable is written inconsistently as c_Text and C_Text, and the expectation subscripts are garbled; please unify notation and typeset the equations cleanly.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the learned instruction is fitted to one training pair and evaluated on 399 held-out pairs, so the headline result is not forced by construction.

full rationale

The central derivation is not circular. The only learned parameter is the text instruction C_Text, optimized on one photo-sketch pair via L_rec (Eq. 3) and L_edit (Eqs. 4-6). The reported SSIM/LPIPS/FID in Table 1 are computed on the remaining 399 pairs of OS-Sketch, so the headline numbers are not the training objective evaluated on the training pair. The optional inference text ('convert the image color to black and white with a white background', Sec. 4.2) is a fixed prompt applied to all test images and is not fit to test labels. The method's components (frozen latent diffusion model, CLIP, and image-prompted editing) are external prior works, not results derived from this paper. The only self-citations (Refs. [49], [51], co-authored by L. Lin) appear in the related-work survey of sketch generation methods and do not supply any load-bearing premise. The Limitation section's caveat about precise sketch textures is a robustness/quality concern, not a circularity step. The absence of error bars and the unspecified choice of the single training pair are legitimate empirical-robustness concerns but do not make the derivation equivalent to its inputs by construction.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The method has no invented physical entities. The main dependences are on a frozen diffusion prior, CLIP embedding alignment, and the one-shot premise. Several hyperparameters and the extra inference prompt are hand-selected and could affect the reported gains.

free parameters (4)
  • Lambda_rec (reconstruction loss weight) = 4
    Chosen by hand in Section 4.1 without ablation or sensitivity analysis.
  • Lambda_edit (edit-direction loss weight) = 0.1
    Chosen by hand in Section 4.1 without sensitivity analysis.
  • Inference additional text prompt = 'convert the image color to black and white with a white background'
    Hand-crafted prompt used during inference in Section 4.2; it performs grayscale conversion and may inflate SSIM/LPIPS against grayscale targets, and it is unclear whether baselines receive the same prompt.
  • Optimization steps N = 12000
    Number of gradient steps for the instruction, set in Section 4.1 without reporting sensitivity.
assumptions (3)
  • domain assumption A pre-trained InstructPix2Pix-style diffusion model can perform face sketch synthesis given an optimized text instruction, with the model frozen.
    Section 3.2, Eq. 3; the method optimizes only C_Text and relies on the diffusion prior to understand sketches.
  • domain assumption The CLIP image embedding difference between photo and sketch lies in the same space as the text embedding C_Text, and cosine similarity is a valid objective for aligning them.
    Section 3.2, Eqs. 4-5; the paper assumes CLIP image and text embeddings are directly comparable.
  • domain assumption A single photo-sketch pair conveys enough information to define a transferable sketch style.
    This is the one-shot premise; Section 4.2 uses one pair for training and tests on the rest.

how reviews work

0 comments
Cite this review

Pith. "Pith review of One-shot Face Sketch Synthesis in the Wild via Generative Diffusion Prior and Instruction Tuning." pith.science (2026). https://pith.science/paper/O2KHKJI4

@misc{pith2026250615312,
  author       = {Pith},
  title        = {Pith review of: One-shot Face Sketch Synthesis in the Wild via Generative Diffusion Prior and Instruction Tuning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O2KHKJI4}},
  note         = {Machine review of arXiv:2506.15312}
}
read the original abstract

Face sketch synthesis is a technique aimed at converting face photos into sketches. Existing face sketch synthesis research mainly relies on training with numerous photo-sketch sample pairs from existing datasets. However, these large-scale discriminative learning methods will have to face problems such as data scarcity and high human labor costs. Once the training data becomes scarce, their generative performance significantly degrades. In this paper, we propose a one-shot face sketch synthesis method based on diffusion models. We optimize text instructions on a diffusion model using face photo-sketch image pairs. Then, the instructions derived through gradient-based optimization are used for inference. To simulate real-world scenarios more accurately and evaluate method effectiveness more comprehensively, we introduce a new benchmark named One-shot Face Sketch Dataset (OS-Sketch). The benchmark consists of 400 pairs of face photo-sketch images, including sketches with different styles and photos with different backgrounds, ages, sexes, expressions, illumination, etc. For a solid out-of-distribution evaluation, we select only one pair of images for training at each time, with the rest used for inference. Extensive experiments demonstrate that the proposed method can convert various photos into realistic and highly consistent sketches in a one-shot context. Compared to other methods, our approach offers greater convenience and broader applicability. The dataset will be available at: https://github.com/HanWu3125/OS-Sketch

Figures

Figures reproduced from arXiv: 2506.15312 by the authors.

Figure 1
Figure 1. An illustration of one-shot sketch synthesis. We call for a solid face sketch model is able to generate [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. A demonstration of Out-of-Distribution (OOD) challenge. When trained with only a single photo-sketch [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Our framework jointly optimizes the text instruction [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: An illustration of the proposed in-the-wild face sketch algorithm. In our framework, only one pair of [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: An illustration of OS-Sketch dataset. Our dataset contains photos with various backgrounds, sexes, [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Qualitative comparisons. We conduct a qualitative comparison of our method with other methods on [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: We exhibit an in-the-wild experiment by using 4 challenging samples selected from the internet for [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Comparison of the generation results of "w/ [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Comparison of the generation results of guidance with hybrid instruction and with solely [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 46 canonical work pages

  1. [1]

    Swapna Agarwal and Dipti Prasad Mukherjee. 2018. Synthesis of realistic facial expressions using expression map. IEEE Transactions on Multimedia21, 4 (2018), 902–914

  2. [2]

    Amir Bar, Yossi Gandelsman, Trevor Darrell, Amir Globerson, and Alexei Efros. 2022. Visual prompting via image inpainting.Advances in Neural Information Processing Systems35 (2022), 25005–25017

  3. [3]

    Tim Brooks, Aleksander Holynski, and Alexei A Efros. 2023. Instructpix2pix: Learning to follow image editing instructions. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 18392–18402

  4. [4]

    Eva Cetinic and James She. 2022. Understanding and creating art with AI: Review and outlook.ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM)18, 2 (2022), 1–22

  5. [5]

    Chaofeng Chen, Xiao Tan, and Kwan-Yee K Wong. 2018. Face sketch synthesis with style transfer using pyramid column feature. In2018 IEEE Winter Conference on Applications of Computer Vision (W ACV). IEEE, 485–493

  6. [6]

    Zijun Deng, Xiangteng He, and Yuxin Peng. 2023. LFR-GAN: local feature refinement based generative adversarial network for text-to-image generation.ACM Transactions on Multimedia Computing, Communications and Applications 19, 6 (2023), 1–18

  7. [7]

    Prafulla Dhariwal and Alexander Nichol. 2021. Diffusion models beat gans on image synthesis.Advances in neural information processing systems34 (2021), 8780–8794

  8. [8]

    Deng-Ping Fan, Ziling Huang, Peng Zheng, Hong Liu, Xuebin Qin, and Luc Van Gool. 2022. Facial-sketch synthesis: A new challenge.Machine Intelligence Research19, 4 (2022), 257–287

Show all 61 references
  1. [9]

    Oran Gafni, Adam Polyak, Oron Ashual, Shelly Sheynin, Devi Parikh, and Yaniv Taigman. 2022. Make-a-scene: Scene-based text-to-image generation with human priors. InEuropean Conference on Computer Vision. Springer, 89–106

  2. [10]

    Fei Gao, Yifan Zhu, Chang Jiang, and Nannan Wang. 2023. Human-Inspired Facial Sketch Synthesis with Dynamic Adaptation. InProceedings of the IEEE/CVF International Conference on Computer Vision. 7237–7247

  3. [11]

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2014. Generative adversarial nets.Advances in neural information processing systems27 (2014)

  4. [12]

    Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. 2022. Prompt-to-prompt image editing with cross attention control.(2022).URL https://arxiv. org/abs/2208.01626(2022)

  5. [13]

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. 2017. Gans trained by a two time-scale update rule converge to a local nash equilibrium.Advances in neural information processing systems 30 (2017)

  6. [14]

    Ziqi Huang, Tianxing Wu, Yuming Jiang, Kelvin CK Chan, and Ziwei Liu. 2023. ReVersion: Diffusion-based relation inversion from images.arXiv preprint arXiv:2303.13495(2023)

  7. [15]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A Efros. 2017. Image-to-image translation with conditional adversarial networks. InProceedings of the IEEE conference on computer vision and pattern recognition. 1125–1134

  8. [16]

    Licheng Jiao, Sibo Zhang, Lingling Li, Fang Liu, and Wenping Ma. 2018. A modified convolutional neural network for face sketch synthesis.Pattern Recognition76 (2018), 125–136

  9. [17]

    Bahjat Kawar, Shiran Zada, Oran Lang, Omer Tov, Huiwen Chang, Tali Dekel, Inbar Mosseri, and Michal Irani. 2023. Imagic: Text-based real image editing with diffusion models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 6007–6017

  10. [18]

    Gwanghyun Kim, Taesung Kwon, and Jong Chul Ye. 2022. Diffusionclip: Text-guided diffusion models for robust image manipulation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2426–2435

  11. [19]

    Bo Li, Xiao Lin, Bin Liu, Zhi-Fen He, and Yu-Kun Lai. 2023. Lightweight text-driven image editing with disentangled content and attributes.IEEE Transactions on Multimedia(2023)

  12. [20]

    Yongkang Li, Qifan Liang, Zhen Han, Wenjun Mai, and Zhongyuan Wang. 2024. Few-shot face sketch-to-photo synthesis via global-local asymmetric image-to-image translation.ACM Transactions on Multimedia Computing, Communications and Applications20, 10 (2024), 1–24

  13. [21]

    Decheng Liu, Jie Li, Nannan Wang, Chunlei Peng, and Xinbo Gao. 2018. Composite components-based face sketch recognition.Neurocomputing302 (2018), 46–54

  14. [22]

    Qingshan Liu, Xiaoou Tang, Hongliang Jin, Hanqing Lu, and Songde Ma. 2005. A nonlinear approach for face sketch synthesis and recognition. In2005 IEEE Computer Society conference on computer vision and pattern recognition (CVPR’05), Vol. 1. IEEE, 1005–1010

  15. [23]

    I Loshchilov. 2017. Decoupled weight decay regularization.arXiv preprint arXiv:1711.05101(2017)

  16. [24]

    Dan Lu, Zhenxue Chen, QM Jonathan Wu, and Xuetao Zhang. 2019. FCN based preprocessing for exemplar-based face sketch synthesis.Neurocomputing365 (2019), 113–124

  17. [25]

    Chenlin Meng, Yang Song, Jiaming Song, Jiajun Wu, Jun-Yan Zhu, and Stefano Ermon. 2021. Sdedit: Image synthesis and editing with stochastic differential equations.arXiv preprint arXiv:2108.01073(2021). J. ACM, Vol. 37, No. 4, Article 111. Publication date: August 2018. 111:16 ...

  18. [26]

    Ron Mokady, Amir Hertz, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. 2023. Null-text inversion for editing real images using guided diffusion models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 6038–6047

  19. [27]

    Thao Nguyen, Yuheng Li, Utkarsh Ojha, and Yong Jae Lee. 2024. Visual instruction inversion: Image editing via image prompting.Advances in Neural Information Processing Systems36 (2024)

  20. [28]

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. 2021. Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741(2021)

  21. [29]

    Lin Nie, Lingbo Liu, Zhengtao Wu, and Wenxiong Kang. 2022. Unconstrained face sketch synthesis via perception- adaptive network and a new benchmark.Neurocomputing494 (2022), 192–202

  22. [30]

    Gaurav Parmar, Krishna Kumar Singh, Richard Zhang, Yijun Li, Jingwan Lu, and Jun-Yan Zhu. 2023. Zero-shot image-to-image translation. InACM SIGGRAPH 2023 Conference Proceedings. 1–11

  23. [31]

    Chunlei Peng, Congyu Zhang, Decheng Liu, Nannan Wang, and Xinbo Gao. 2023. HiFiSketch: High Fidelity Face Photo-Sketch Synthesis and Manipulation.IEEE Transactions on Image Processing(2023)

  24. [32]

    Chikontwe Philip and Lee Hyo Jong. 2017. Face sketch synthesis using conditional adversarial networks. In2017 International Conference on Information and Communication Technology Convergence (ICTC). IEEE, 373–378

  25. [33]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. InInternational conference on machine learnin...

  26. [34]

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. 2022. High-resolution image synthesis with latent diffusion models. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 10684–10695

  27. [35]

    Chitwan Saharia, William Chan, Huiwen Chang, Chris Lee, Jonathan Ho, Tim Salimans, David Fleet, and Mohammad Norouzi. 2022. Palette: Image-to-image diffusion models. InACM SIGGRAPH 2022 conference proceedings. 1–10

  28. [36]

    Bowen Sun, Guo Lu, and Shibao Zheng. 2023. DiFace: Cross-Modal Face Recognition through Controlled Diffusion. ACM Transactions on Multimedia Computing, Communications and Applications(2023)

  29. [37]

    Hamidou Tembine. 2019. Deep learning meets game theory: Bregman-based algorithms for interactive deep generative adversarial networks.IEEE transactions on cybernetics50, 3 (2019), 1132–1145

  30. [38]

    Jiahang Tu, Wei Ji, Hanbin Zhao, Chao Zhang, Roger Zimmermann, and Hui Qian. 2025. Driveditfit: Fine-tuning diffusion transformers for autonomous driving data generation.ACM Transactions on Multimedia Computing, Commu- nications and Applications21, 3 (2025), 1–29

  31. [39]

    Narek Tumanyan, Michal Geyer, Shai Bagon, and Tali Dekel. 2023. Plug-and-play diffusion features for text-driven image-to-image translation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 1921–1930

  32. [40]

    Nannan Wang, Wenjin Zha, Jie Li, and Xinbo Gao. 2018. Back projection: An effective postprocessing method for GAN-based face sketch synthesis.Pattern Recognition Letters107 (2018), 59–65

  33. [41]

    Nannan Wang, Mingrui Zhu, Jie Li, Bin Song, and Zan Li. 2017. Data-driven vs. model-driven: Fast face sketch synthesis. Neurocomputing257 (2017), 214–221

  34. [42]

    Xiaogang Wang and Xiaoou Tang. 2008. Face photo-sketch synthesis and recognition.IEEE transactions on pattern analysis and machine intelligence31, 11 (2008), 1955–1967

  35. [43]

    Xinlong Wang, Wen Wang, Yue Cao, Chunhua Shen, and Tiejun Huang. 2023. Images speak in images: A generalist painter for in-context visual learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 6830–6839

  36. [44]

    Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. 2004. Image quality assessment: from error visibility to structural similarity.IEEE transactions on image processing13, 4 (2004), 600–612

  37. [45]

    Yuxin Wen, Neel Jain, John Kirchenbauer, Micah Goldblum, Jonas Geiping, and Tom Goldstein. 2024. Hard prompts made easy: Gradient-based discrete optimization for prompt tuning and discovery.Advances in Neural Information Processing Systems36 (2024)

  38. [46]

    Xian Wu, Kun Xu, and Peter Hall. 2017. A survey of image synthesis and editing with generative adversarial networks. Tsinghua Science and Technology22, 6 (2017), 660–674

  39. [47]

    Ran Yi, Yong-Jin Liu, Yu-Kun Lai, and Paul L Rosin. 2019. Apdrawinggan: Generating artistic portrait drawings from face photos with hierarchical gans. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 10743–10752

  40. [48]

    Jun Yu, Xingxin Xu, Fei Gao, Shengjie Shi, Meng Wang, Dacheng Tao, and Qingming Huang. 2020. Toward realistic face photo–sketch synthesis via composition-aided GANs.IEEE transactions on cybernetics51, 9 (2020), 4350–4362

  41. [49]

    Dongyu Zhang, Liang Lin, Tianshui Chen, Xian Wu, Wenwei Tan, and Ebroul Izquierdo. 2016. Content-adaptive sketch portrait generation by decompositional representation learning.IEEE transactions on image processing26, 1 (2016), J. ACM, Vol. 37, No. 4, Article 111. Publication d...

  42. [50]

    Jiewei Zhang, Nannan Wang, Xinbo Gao, Dacheng Tao, and Xuelong Li. 2011. Face sketch-photo synthesis based on support vector regression. In2011 18th IEEE International Conference on Image Processing. IEEE, 1125–1128

  43. [51]

    Liliang Zhang, Liang Lin, Xian Wu, Shengyong Ding, and Lei Zhang. 2015. End-to-end photo-sketch generation via fully convolutional representation learning. InProceedings of the 5th ACM on International Conference on Multimedia Retrieval. 627–634

  44. [52]

    Mingjin Zhang, Nannan Wang, Yunsong Li, and Xinbo Gao. 2019. Deep latent low-rank representation for face sketch synthesis.IEEE transactions on neural networks and learning systems30, 10 (2019), 3109–3123

  45. [53]

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. 2018. The unreasonable effectiveness of deep features as a perceptual metric. InProceedings of the IEEE conference on computer vision and pattern recognition. 586–595

  46. [54]

    Wei Zhang, Xiaogang Wang, and Xiaoou Tang. 2011. Coupled information-theoretic encoding for face photo-sketch recognition. InCVPR 2011. IEEE, 513–520

  47. [55]

    Mohan Zhou, Yalong Bai, Qing Yang, and Tiejun Zhao. 2024. StyleInject: Parameter Efficient Tuning of Text-to-Image Diffusion Models.ACM Transactions on Multimedia Computing, Communications and Applications(2024)

  48. [56]

    Mingrui Zhu, Jie Li, Nannan Wang, and Xinbo Gao. 2019. A deep collaborative framework for face photo–sketch synthesis.IEEE transactions on neural networks and learning systems30, 10 (2019), 3096–3108

  49. [57]

    Mingrui Zhu and Nannan Wang. 2016. A simple and fast method for face sketch synthesis. InProceedings of the International Conference on Internet Multimedia Computing and Service. 168–171

  50. [58]

    Mingrui Zhu, Nannan Wang, Xinbo Gao, and Jie Li. 2017. Deep graphical feature learning for face sketch synthesis. In Proceedings of the 26th international joint conference on artificial intelligence. 3574–3580

  51. [59]

    Mingrui Zhu, Zicheng Wu, Nannan Wang, Heng Yang, and Xinbo Gao. 2023. Dual conditional normalization pyramid network for face photo-sketch synthesis.IEEE Transactions on Circuits and Systems for Video Technology33, 9 (2023), 5200–5211

  52. [60]

    Zhu, Mingrui and Li, Jie and Wang, Nannan and Gao, Xinbo. 2020. Knowledge distillation for face photo–sketch synthesis.IEEE Transactions on Neural Networks and Learning Systems33, 2 (2020), 893–906

  53. [61]

    Kaifeng Zou, Sylvain Faisan, Boyang Yu, Sébastien Valette, and Hyewon Seo. 2024. 4d facial expression diffusion model. ACM Transactions on Multimedia Computing, Communications and Applications21, 1 (2024), 1–23. Received 20 February 2007; revised 12 March 2009; accepted 5 June...

Pith tools

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