Pith. sign in

REVIEW 4 major objections 5 minor 32 references

Beyond Facial Consistency: Personalized Person Image Generation with Holistic Identity Preservation

T0 review · 4 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read The paper claims that holistic identity preservation in personalized person image generation reduces to coordinating a face branch and an appearance branch with time-aware scalar gates and region-aware supervision, without retraining the di

desk verdict Useful engineering note on combining a face branch and an appearance branch for person personalization; the reported gains are small and the 'better trade-off' claim needs a defined criterion and open artifacts before it convinces me. read the letter →

arxiv 2607.25622 v1 pith:5G6SZKU3 submitted 2026-07-28 cs.CV

Yuxuan Xiao , Shanshan Zhang , Jian Yang , Shengcai Liao This is my paper · ORCID
classification cs.CV
keywords personalizedimagegenerationidentitypreservationfaceconsistencyappearancediffusiontransformertemporalgatingregion-awareoptimizationholistic
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 establish that the usual trade-off in personalized person image generation—keeping either the face or the overall appearance consistent, but not both—is not a hard limit of the generation model but a coordination problem between two conditioning pathways. It proposes a simple two-branch setup, one branch carrying global appearance cues and one carrying facial identity, and then shows that instability in that setup comes from the face branch dominating the prediction. The proposed fix, DBS, uses timestep-dependent scalar gates to rebalance the branches during denoising and a region-aware loss that protects facial identity while letting appearance guidance strengthen. If correct, this makes holistic identity preservation a lightweight fine-tuning problem on a frozen diffusion transformer, and the new Pexels-100 benchmark gives the community a shared way to measure it.

What carries the argument

Adaptive Temporal Gating (ATG): per-layer scalar gates that are functions of the diffusion timestep; the face gate is 1 + delta*tanh(...) to stay close to its pretrained regime, and the appearance gate is a time-dependent multiple of the face gate with a progressive relaxation coefficient that starts near zero and grows with training. These gates multiply the branch residuals before injection into the transformer, letting the model shift weight from face to appearance as denoising proceeds. Region-Aware Optimization (RAO): the denoising error map is split into face-mask, person-mask, and global regions; each regional loss is normalized by an EMA reference, and a log-sum-exp aggregation with

What would settle it

Concretely, measure the perturbation gap after removing each branch, as the paper does, and then reimplement the face branch with the same injection mechanism and feature dimensions as the appearance branch while retraining only the gates; if face over-dominance remains, relative signal strength is not the cause and the gate hypothesis fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that face fidelity and appearance consistency can be jointly preserved by explicitly separating identity into two conditioning branches and then coordinating them with scalar, time-dependent gates plus spatially structured supervision. The face branch is treated as a stable anchor: its gate is bounded near unity, and the appearance branch's gate is initialized relative to the face gate and gradually relaxed during training. A Pareto-style region-aware loss then activates facial and appearance objectives only when they underperform relative to a running average, while the global objective continues to optimize image quality. The authors report th

Load-bearing premise

The load-bearing premise is that branch conflict is caused by relative signal strength, not by incompatible representation spaces or injection mechanisms; if conflict persists under equalized architectures, the scalar gates cannot fix it.

Editorial extensions

If this is right

  • Face fidelity and appearance consistency can be improved jointly by tuning only gate parameters, leaving the generative backbone frozen.
  • The naive dual-branch setup plus DBS provides a controllable basic framework: users can strengthen or weaken face versus appearance conditioning by adjusting gate scales.
  • The perturbation-gap measurement quantifies branch dominance and predicts where coordination is needed.
  • Region-aware Pareto-style supervision protects identity-critical regions without sacrificing global image quality.
  • The Pexels-100 benchmark offers a reusable protocol for comparing holistic identity rather than face-only fidelity.

Reading between the lines

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

  • If the rebalancing is truly achieved by scalar gates, the same recipe should transfer to other heterogeneous conditioning branches (pose, style, layout) on frozen diffusion transformers—a testable extension the paper does not run.
  • Because only gate parameters are trained, DBS-like tuning could be an inexpensive adapter layer for new backbone versions, letting identity preservation ride on top of improved generators.
  • The paper's diagnosis that face-branch over-dominance reflects injection-mechanism strength rather than inherent importance implies that aligning injection mechanisms (same pathway type) might achieve similar balance without gates—an alternative the authors do not compare.
  • The authors concede that complex poses and fine structures like hands still distort; whether gating can ever address structural defects is an open question beyond the paper's current scope.
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

4 major / 5 minor

Summary. The paper addresses holistic identity preservation in personalized person image generation, proposing a naive dual-branch (NDB) baseline that combines a global appearance branch (InstantCharacter-style) and a local face branch (InfiniteYou-style) on the FLUX.1-dev backbone. It then introduces Dynamic Balancing Scaling (DBS), a fine-tuning strategy with two components: Adaptive Temporal Gating (ATG), which applies time-dependent scalar gates to the face and appearance branch residuals (Eqs. 2-5), and Region-Aware Optimization (RAO), which decomposes the denoising loss into face, appearance, and global terms with a Pareto-style soft aggregation (Eqs. 6-10). The paper also introduces a new benchmark, Pexels-100, and reports quantitative metrics (Face, ReID, DINOv2, CLIP-T) and qualitative comparisons. The central claim is that DBS achieves a better trade-off between facial fidelity and appearance consistency than existing open-source baselines.

Significance. If validated, the work offers a lightweight, backbone-frozen fine-tuning recipe for coordinating heterogeneous identity branches in a diffusion transformer, which could be practically useful and generalizable beyond person generation. The introduction of a dedicated holistic-consistency benchmark is also a potentially valuable community resource. However, the empirical evidence in the current manuscript is substantially weaker than the claims. All quantitative results come from a single run with no error bars or significance tests; the Pexels-100 benchmark is introduced but not released; and the central 'better trade-off' claim is not operationalized with any holistic criterion, preference study, or Pareto analysis. The method's mechanism is plausible but rests on the assumption that scalar, time-dependent gates can fix branch imbalance, which is asserted rather than demonstrated. The paper ships no code, no benchmark, and no machine-checked artifacts; its current value is mainly conceptual.

major comments (4)
  1. [Abstract / Table 1] The central claim 'DBS achieves a better trade-off' is not operationalized and, as stated, is unfalsifiable. In Table 1, DBS does not Pareto-dominate open-source baselines: Qwen-Image has higher ReID (0.9230 vs 0.9122) and CLIP-T (0.2279 vs 0.2008) but lower Face (0.4640 vs 0.7875); InfiniteYou has lower ReID but higher Face. Whether DBS's point is preferable is a value judgment. The paper needs a defined trade-off criterion, a combined holistic metric, a preference study, or an explicit Pareto-frontier analysis. Without this, the main abstract and conclusion claims are not testable.
  2. [Tables 2 and 3, Sections 4.3-4.4] All quantitative results are single-run with no error bars, seeds, or significance tests. In Table 2, the full model differs from NDB by Face +0.013, ReID +0.002, DINOv2 +0.018, and CLIP-T +0.002; these differences are within typical run-to-run noise. The same issue affects Table 3 comparisons with MSE, LS, and FairGrad. The ablations therefore do not currently support the claim that ATG and RAO contribute meaningfully. Report means and standard deviations over at least three independent runs and, if appropriate, paired significance tests.
  3. [Section 4.2 / Pexels-100] The evaluation is built on a self-constructed benchmark that is not released, and the construction is described in only one sentence: 'We used ChatGPT to generate 100 text prompts and paired them with these images.' No details are given on the number of identities, the prompt-image pairing procedure, filtering criteria, or diversity. The ReID metric (TransMatcher) is co-authored by a co-author, creating a conflict-of-interest concern that should be addressed with independent metrics or human evaluation. The paper should release the benchmark (or a clear description with statistics) and include additional holistic metrics (e.g., CLIP appearance similarity, human preference study).
  4. [Section 3.3 and Figure 2] The motivation for ATG rests on the perturbation analysis in Figure 2, which reports an average difference in noise-prediction change when removing each branch. No details are given about how this average is computed, across how many images, or with what variance. Moreover, the proposed scalar gates (Eqs. 2-3) only rescale branch residual magnitudes; if the instability arises from incompatible representation spaces (cross-attention residuals vs ControlNet residuals), such gating cannot address it. The paper should provide quantitative evidence that the gates actually rebalance branch contributions over the denoising trajectory, e.g., by plotting the learned gate values and the resulting perturbation magnitudes before/after training.
minor comments (5)
  1. [Section 3.4 (first paragraph)] Typo: 'the denoising trajectory' should be 'along the denoising trajectory'. Also, 'at different identity information' should be 'at different identity granularities' or similar.
  2. [Figure 3 caption] Abbreviation inconsistency: the caption defines 'RAB: Region Aware Optimization' but the text and equations use 'RAO'. Please unify.
  3. [Table 1] The row label 'Upcoming Ours' is awkward and perhaps an artifact; use 'Ours'. Also, the names 'Nano banana' and 'Nano banana2' should be capitalized consistently.
  4. [Section 4.2] The face similarity threshold of 0.75 for filtering PPR10K is not attributed to a specific face recognition model. Specify which model and which image pairs were used to compute similarity.
  5. [Figure 4] Only Face scores are shown per generated image. Since holistic consistency is the focus, include ReID or DINOv2 scores for each qualitative example, or at least state how the Face scores were computed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the method is empirical and no equation or fitted parameter reduces the claimed result to its own inputs.

full rationale

I walked the derivation chain from the naive dual-branch baseline through Adaptive Temporal Gating (Eqs. 1-5) and Region-Aware Optimization (Eqs. 6-10). The gate parameters are learned from training data and are not fitted to the evaluation metrics; the RAO loss uses EMA self-referencing only as an optimization weighting, not as a definition of the target. The Pexels-100 benchmark is constructed by the authors and the ReID metric (TransMatcher) is co-authored by one of the authors, but neither embeds the method's outputs or parameters, and no specific reduction of the central claim to these choices is exhibited. The paper's 'better trade-off' claim is under-specified (no Pareto-dominance or preference study), but that is a falsifiability/correctness concern, not a circularity. No step meets the bar of showing Eq. X = Eq. Y by construction, a fitted parameter renamed as a prediction, or a load-bearing self-citation chain.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The central claim is empirical; no derivation is presented. The main free parameters are gate ranges, RAO temperature/EMA, and the face-similarity filter threshold; none are derived from first principles or sensitivity-analyzed.

free parameters (8)
  • δ (face gate range) = not reported
    Bounds the face-branch gate around 1 in Eq. 2; chosen by hand, controls how much the facial branch can adapt.
  • γ (appearance gate deviation) = not reported
    Controls deviation range of the appearance gate in Eq. 3; hand-set.
  • α(τ) (progressive relaxation coefficient) = not reported
    Schedule for relaxing the appearance branch toward independence; not specified.
  • η (RAO temperature) = 0.1
    Softmax temperature in Eq. 10; hand-set.
  • μ (EMA decay in RAO) = not reported
    Exponential moving average decay for reference losses in Eq. 7; hand-set.
  • ϵ (RAO normalization constant) = not reported
    Stability constant in Eq. 8.
  • s_a (appearance residual scale) = not reported
    Scaling factor applied to appearance attention residuals in Eq. 5; no value given.
  • Face similarity threshold for data filtering = 0.75
    Used to construct same-identity training pairs from PPR10K; selection criterion not justified.
assumptions (5)
  • domain assumption FLUX.1-dev is a valid generative backbone for this task.
    Used throughout; no re-training or verification of base model behavior.
  • domain assumption Appearance and face branch residuals can be additively injected into DiT hidden states.
    Assumes compatibility of cross-attention and ControlNet-style conditioning (Eq. 4-5).
  • domain assumption Removing one branch and measuring noise-prediction change is a valid proxy for branch dominance.
    Motivates the whole method (Sec. 1, Fig. 2).
  • domain assumption PPR10K can be filtered into same-identity pairs using a face-similarity threshold.
    Training data construction (Sec. 4.2); relies on dataset group-level consistency and threshold adequacy.
  • domain assumption ArcFace, TransMatcher ReID, and DINOv2 similarities measure holistic identity.
    Primary evidence for the central claim; no human evaluation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Beyond Facial Consistency: Personalized Person Image Generation with Holistic Identity Preservation." pith.science (2026). https://pith.science/paper/5G6SZKU3

@misc{pith2026260725622,
  author       = {Pith},
  title        = {Pith review of: Beyond Facial Consistency: Personalized Person Image Generation with Holistic Identity Preservation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5G6SZKU3}},
  note         = {Machine review of arXiv:2607.25622}
}
read the original abstract

Personalized person image generation requires preserving subject identity across both local facial details and broader appearance cues. Existing methods typically emphasize only one level of identity information, leading to an inherent trade-off between facial fidelity and overall appearance consistency. To address this, we first propose a simple dual-branch baseline that unifies global appearance control and local facial control within a shared generation framework. This simple combination of different branches yields promising results, but suffers from instability in practice due to uncoordinated branch contributions. To this end, we propose Dynamic Balancing Scaling (DBS), a fine-tuning strategy for improving face and appearance identity coordination. DBS consists of two components: adaptive temporal gating, which dynamically modulates branch contributions along the denoising trajectory, and region-aware optimization, which improves the coordination of facial, appearance, and global supervision. Together, these designs alleviate persistent face-branch over-dominance and encourage more effective appearance-aware guidance. We also introduce Pexels-100, a benchmark for evaluating holistic identity consistency in personalized person generation. Experiments show that DBS achieves a better trade-off between facial fidelity and appearance consistency than existing open-source baselines, while providing a controllable basic framework for holistic identity modeling.

Figures

Figures reproduced from arXiv: 2607.25622 by the authors.

Figure 1
Figure 1. Comparison of our method with subject consistency and face consistency methods. Please [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Our method improves appearance consistency while maintaining facial stability by coordi [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Overview of our framework. Built upon a naive dual-branch architecture, DBS introduces [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison across different methods. “NDB” stands for Naive Dual-Branch. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Modeling complex human poses or fine-grained structures remains challenging. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 6 linked inside Pith

  1. [1]

    Qwen3-vl technical report.arXiv preprint arXiv:2511.21631, 2025

    Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, et al. Qwen3-vl technical report.arXiv preprint arXiv:2511.21631, 2025

    arXiv 2025

  2. [2]

    FLUX.1 Release: Announcing Black Forest Labs

    Black Forest Labs. FLUX.1 Release: Announcing Black Forest Labs. https:// blackforestlabs.ai/announcing-black-forest-labs/, 2024. Accessed: 2026-04-30

    2024

  3. [3]

    Arcface: Additive angular margin loss for deep face recognition

    Jiankang Deng, Jia Guo, Niannan Xue, and Stefanos Zafeiriou. Arcface: Additive angular margin loss for deep face recognition. InCVPR, 2019

    2019

  4. [4]

    Introducing nano banana pro

    Google. Introducing nano banana pro. https://blog.google/innovation-and-ai/ products/nano-banana-pro/, 2025. Accessed: 2026-04-30

    2025

  5. [5]

    Nano banana 2: Google’s latest ai image generation model, 2026

    Google. Nano banana 2: Google’s latest ai image generation model, 2026. Accessed: 2026-04- 30

    2026

  6. [6]

    Pulid: Pure and lightning id customization via contrastive alignment.NeurIPS, 2024

    Zinan Guo, Yanze Wu, Zhuowei Chen, Lang Chen, Peng Zhang, and Qian He. Pulid: Pure and lightning id customization via contrastive alignment.NeurIPS, 2024

    2024

  7. [7]

    Clipscore: A reference-free evaluation metric for image captioning

    Jack Hessel, Ari Holtzman, Maxwell Forbes, Ronan Le Bras, and Yejin Choi. Clipscore: A reference-free evaluation metric for image captioning. InEMNLP, 2021

    2021

  8. [8]

    Consistentid: Portrait generation with multimodal fine-grained identity preserving.IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(5):5639–5654, 2026

    Jiehui Huang, Xiao Dong, Wenhui Song, Zheng Chong, Zhenchao Tang, Jun Zhou, Yuhao Cheng, Long Chen, Hanhui Li, Yiqiang Yan, Shengcai Liao, and Xiaodan Liang. Consistentid: Portrait generation with multimodal fine-grained identity preserving.IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(5):5639–5654, 2026. 10

    2026

Show all 32 references
  1. [9]

    Joyai-image: A unified framework for image generation and editing

    JD OpenSource. Joyai-image: A unified framework for image generation and editing. https: //github.com/jd-opensource/JoyAI-Image, 2025. Accessed: 2026-04-30

    2025

  2. [10]

    Infiniteyou: Flexible photo recrafting while preserving your identity

    Liming Jiang, Qing Yan, Yumin Jia, Zichuan Liu, Hao Kang, and Xin Lu. Infiniteyou: Flexible photo recrafting while preserving your identity. InICCV, 2025

    2025

  3. [11]

    Black Forest Labs, Stephen Batifol, Andreas Blattmann, Frederic Boesel, Saksham Consul, Cyril Diagne, Tim Dockhorn, Jack English, Zion English, Patrick Esser, et al. Flux. 1 kontext: Flow matching for in-context image generation and editing in latent space.arXiv preprint arXiv...

    2025 arXiv

  4. [12]

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. InICML, 2023

    2023

  5. [13]

    Photomaker: Customizing realistic human photos via stacked id embedding

    Zhen Li, Mingdeng Cao, Xintao Wang, Zhongang Qi, Ming-Ming Cheng, and Ying Shan. Photomaker: Customizing realistic human photos via stacked id embedding. InCVPR, 2024

    2024

  6. [14]

    Ppr10k: A large-scale portrait photo retouching dataset with human-region mask and group-level consistency

    Jie Liang, Hui Zeng, Miaomiao Cui, Xuansong Xie, and Lei Zhang. Ppr10k: A large-scale portrait photo retouching dataset with human-region mask and group-level consistency. In CVPR, 2021

    2021

  7. [15]

    Transmatcher: Deep image matching through transformers for generalizable person re-identification.NeurIPS, 2021

    Shengcai Liao and Ling Shao. Transmatcher: Deep image matching through transformers for generalizable person re-identification.NeurIPS, 2021

    2021

  8. [16]

    Decoupled weight decay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. InICLR, 2019

    2019

  9. [17]

    Visual persona: Foundation model for full-body human customization

    Jisu Nam, Soowon Son, Zhan Xu, Jing Shi, Difan Liu, Feng Liu, Seungryong Kim, and Yang Zhou. Visual persona: Foundation model for full-body human customization. InCVPR, 2025

    2025

  10. [18]

    Chatgpt.https://chat.openai.com/, 2024

    OpenAI. Chatgpt.https://chat.openai.com/, 2024. Accessed: 2026-04-30

    2024

  11. [19]

    Gpt-image-1.5

    OpenAI. Gpt-image-1.5. https://platform.openai.com/docs/guides/images, 2024. Accessed: 2026-04-30

    2024

  12. [20]

    Gpt-image-2

    OpenAI. Gpt-image-2. https://platform.openai.com/docs/guides/images, 2025. Accessed: 2026-04-30

    2025

  13. [21]

    Dinov2: Learning robust visual features without supervision.Transactions on Machine Learning Research Journal, 2024

    Maxime Oquab, Timothée Darcet, Théo 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.Transactions on Machine Learning Research Journal, 2024

    2024

  14. [22]

    Ominicontrol: Minimal and universal control for diffusion transformer

    Zhenxiong Tan, Songhua Liu, Xingyi Yang, Qiaochu Xue, and Xinchao Wang. Ominicontrol: Minimal and universal control for diffusion transformer. InICCV, 2025

    2025

  15. [23]

    Instantcharacter: Personalize any characters with a scalable diffusion transformer framework.arXiv preprint arXiv:2504.12395, 2025

    Jiale Tao, Yanbing Zhang, Qixun Wang, Yiji Cheng, Haofan Wang, Xu Bai, Zhengguang Zhou, Ruihuang Li, Linqing Wang, Chunyu Wang, et al. Instantcharacter: Personalize any characters with a scalable diffusion transformer framework.arXiv preprint arXiv:2504.12395, 2025

    2025 arXiv

  16. [24]

    Instantid: Zero-shot identity-preserving generation in seconds.arXiv preprint arXiv:2401.07519, 2024

    Qixun Wang, Xu Bai, Haofan Wang, Zekui Qin, Anthony Chen, Huaxia Li, Xu Tang, and Yao Hu. Instantid: Zero-shot identity-preserving generation in seconds.arXiv preprint arXiv:2401.07519, 2024

    2024 arXiv

  17. [25]

    Fairhuman: Boosting hand and face quality in human image generation with minimum potential delay fairness in diffusion models

    Yuxuan Wang, Tianwei Cao, Huayu Zhang, Zhongjiang He, Kongming Liang, and Zhanyu Ma. Fairhuman: Boosting hand and face quality in human image generation with minimum potential delay fairness in diffusion models. InICCV, 2025

    2025

  18. [26]

    Qwen-image technical report.arXiv preprint arXiv:2508.02324, 2025

    Chenfei Wu, Jiahao Li, Jingren Zhou, Junyang Lin, Kaiyuan Gao, Kun Yan, Sheng-ming Yin, Shuai Bai, Xiao Xu, Yilei Chen, et al. Qwen-image technical report.arXiv preprint arXiv:2508.02324, 2025

    2025 arXiv

  19. [27]

    Less-to-more generalization: Unlocking more controllability by in-context generation

    Shaojin Wu, Mengqi Huang, Wenxu Wu, Yufeng Cheng, Fei Ding, and Qian He. Less-to-more generalization: Unlocking more controllability by in-context generation. InICCV, 2025. 11

    2025

  20. [28]

    Ip-adapter: Text compatible image prompt adapter for text-to-image diffusion models.arXiv preprint arXiv:2308.06721, 2023

    Hu Ye, Jun Zhang, Sibo Liu, Xiao Han, and Wei Yang. Ip-adapter: Text compatible image prompt adapter for text-to-image diffusion models.arXiv preprint arXiv:2308.06721, 2023

    2023 arXiv

  21. [29]

    Sigmoid loss for language image pre-training

    Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. Sigmoid loss for language image pre-training. InICCV, 2023

    2023

  22. [30]

    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. InICCV, 2023

    2023

  23. [31]

    Flashface: Human image personalization with high-fidelity identity preservation.arXiv preprint arXiv:2403.17008, 2024

    Shilong Zhang, Lianghua Huang, Xi Chen, Yifei Zhang, Zhi-Fan Wu, Yutong Feng, Wei Wang, Yujun Shen, Yu Liu, and Ping Luo. Flashface: Human image personalization with high-fidelity identity preservation.arXiv preprint arXiv:2403.17008, 2024

    2024

  24. [32]

    Easycontrol: Adding efficient and flexible control for diffusion transformer

    Yuxuan Zhang, Yirui Yuan, Yiren Song, Haofan Wang, and Jiaming Liu. Easycontrol: Adding efficient and flexible control for diffusion transformer. InICCV, 2025. 12

    2025

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.