CoChat
Sign Up
Cited by 48Generative Adversarial Networks and Image SynthesisAdvanced Neural Network ApplicationsMultimodal Machine Learning Applications

The DOI matches its Crossref record; 2 of the 2 references it cites were found in a registry or an index.

OminiControl: Minimal and Universal Control for Diffusion Transformer

Open access
Tan, Zhenxiong, Liu, Songhua, Yang, Xingyi, Xue, Qiaochu, Wang, Xinchao·2025·2025 IEEE/CVF International Conference on Computer Vision (ICCV)
DOI 10.1109/iccv51701.2025.01386
Open this paper in CoChat — free accountOpen PDF

Abstract

We present OminiControl, a novel approach that rethinks how image conditions are integrated into Diffusion Transformer (DiT) architectures. Current image conditioning methods either introduce substantial parameter overhead or handle only specific control tasks effectively, limiting their practical versatility. OminiControl addresses these limitations through three key innovations: (1) a minimal architectural design that leverages the DiT's own VAE encoder and transformer blocks, requiring just 0.1% additional parameters; (2) a unified sequence processing strategy that combines condition tokens with image tokens for flexible token interactions; and (3) a dynamic position encoding mechanism that adapts to both spatially-aligned and non-aligned control tasks. Our extensive experiments show that this streamlined approach not only matches but surpasses the performance of specialized methods across multiple conditioning tasks. To overcome data limitations in subject-driven generation, we also introduce Subjects200K, a large-scale dataset of identity-consistent image pairs synthesized using DiT models themselves. This work demonstrates that effective image control can be achieved without architectural complexity, opening new possibilities for efficient and versatile image generation systems.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/iccv51701.2025.01386

Cites

2 of 2 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Image customization approaches that reason over reference and noise latents tend to rely on implicitly learned spatial cues, limiting spatial positioning.

    “However, these approaches tend to rely on implicitly learned spatial cues, limiting spatial positioning.”

    From DreamingComics: A Story Visualization Pipeline via Subject and Layout Customized Generation using Video Models (2025) · cites this paper as [52]

    Section 3.2.2 (p. 4) reports that shared position indices for condition and noisy-image tokens can impair non-aligned, subject-driven generation. Section 2.2 (p. 2) also describes limits of existing methods on spatially non-aligned tasks. Neither passage says these approaches rely on implicitly learned spatial cues or establishes the claimed general limitation on spatial positioning. (LLM, conf=0.76)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 61

  1. 10.1109/CVPR52688.2022.01767 · DOI
  2. Bar-Tal, 2023, Multidiffusion: Fusing diffusion paths for controlled image generation
  3. 10.1109/cvpr52734.2025.01718 · DOI
  4. Chen, 2023, Pixart- $\alpha$: Fast training of diffusion transformer for photorealistic text-to-image synthesis, arXiv preprint
  5. Chen, 2024, Pixart- $\{\backslash \operatorname{delta}\}$: Fast and controllable image generation with latent consistency models, arXiv preprint
  6. Choi, 2024, Style-friendly snr sampler for style-driven generation, arXiv preprint
  7. 2024, Peak signal-to-noise ratio
  8. 10.1109/ICECA.2018.8474715 · DOI
  9. Esser, 2024, Scaling rectified flow transformers for high-resolution image synthesis
  10. Gal, 2022, An image is worth one word: Personalizing text-to-image generation using textual inversion, arXiv preprint
  11. 10.1145/3422622 · DOI · Paper page
  12. Han, 2024, Ace: All-round creator and editor following instructions via diffusion transformer, arXiv preprint
  13. Hate, 2024, Text-to-image-2m dataset
  14. He, 2025, Anystory: Towards unified single and multiple subject personalization in text-to-image generation
  15. He, 2024, Dynamiccontrol: Adaptive condition selection for improved text-to-image generation
  16. Heusel, 2017, Gans trained by a two time-scale update rule converge to a local nash equilibrium, Advances in neural information processing systems, 30
  17. Ho, 2020, Denoising diffusion probabilistic models, Advances in neural information processing systems, 33, 6840
  18. Hua, 2023, Dreamtuner: Single image is enough for subject-driven generation
  19. Huang, 2024, Group diffusion transformers are unsupervised multitask learners
  20. Huang, 2024, In-context lora for diffusion transformers, arXiv preprint
  21. 10.1109/ICCV48922.2021.00510 · DOI
  22. Kingma, 2013, Auto-encoding variational bayes, arXiv preprint
  23. 10.1109/CVPR52729.2023.00192 · DOI · Paper page
  24. 2024, Flux: Official inference repository for flux. 1 models
  25. 2024, Flux.1-dev-controlnet-union-pro
  26. 10.52202/075280-1312 · DOI
  27. 10.52202/075280-1312 · DOI
  28. 10.1109/CVPR52733.2024.00825 · DOI
  29. Li, 2025, Controlar: Controllable image generation with autoregressive models
  30. 10.52202/079017-4095 · DOI
  31. Lipman, 2022, Flow matching for generative modeling, arXiv preprint
  32. 10.1145/3641519.3657469 · DOI
  33. Ma, 2024, Character-adapter: Prompt-guided region control for highfidelity character customization
  34. 10.1109/ICCVW69036.2025.00207 · DOI
  35. Meng, 2021, Sdedit: Guided image synthesis and editing with stochastic differential equations, arXiv preprint
  36. Mishchenko, 2024, Prodigy: An expeditiously adaptive parameter-free learner
  37. 10.1609/aaai.v38i5.28226 · DOI
  38. Pan, 2023, Kosmos-g: Generating images in context with multimodal large language models, arXiv preprint
  39. Pan, 2020, Multimodal attention for speech emotion recognition, arXiv preprint
  40. Podell, 2023, Sdxl: Improving latent diffusion models for high-resolution image synthesis, arXiv preprint
  41. Qin, 2023, Unicontrol: A unified diffusion model for controllable visual generation in the wild, arXiv preprint
  42. Radford, 2021, Learning transferable visual models from natural language supervision, 8748
  43. 10.1109/cvpr52688.2022.01042 · DOI · Paper page
  44. Ronneberger, 2015, Unet: Convolutional networks for biomedical image segmentation, 234
  45. Rout, 2024, Semantic image inversion and editing using rectified stochastic differential equations, arXiv preprint
  46. 10.1109/CVPR52729.2023.02155 · DOI · Paper page
  47. 10.1145/3528233.3530757 · DOI · Paper page
  48. 10.52202/068431-2643 · DOI · Paper page
  49. 10.1109/CVPR52733.2024.00816 · DOI
  50. 10.1007/978-3-031-73247-8_6 · DOI
  51. 10.1609/aaai.v37i2.25353 · DOI
  52. Wang, 2024, Instantid: Zero-shot identity-preserving generation in seconds, arXiv preprint
  53. 10.1109/TIP.2003.819861 · DOI · Paper page
  54. 10.1109/ICCV51070.2023.01461 · DOI
  55. 10.1007/s11263-024-02227-z · DOI
  56. 2024, Flux-ip-adapter
  57. 10.1109/CVPRW56347.2022.00126 · DOI
  58. Ye, 2023, Ipadapter: Text compatible image prompt adapter for text-to-image diffusion models, arXiv preprint
  59. 10.1109/ICCV51070.2023.00355 · DOI · Paper page
  60. 10.1109/CVPR52733.2024.00771 · DOI
  61. 10.52202/075280-0491 · DOI

Cited by

Sorted by citations

Metadata

Authors
Tan, Z., Liu, S., Yang, X., Xue, Q., & Wang, X.
Affiliations
  • National University of Singapore (SG)
  • National University of Singapore
ORCID
Venue
2025 IEEE/CVF International Conference on Computer Vision (ICCV)
Publisher
IEEE
Published
2025-10-19
Pages
14940-14950
DOI
10.1109/iccv51701.2025.01386
OpenAlex
W4405032013
Cited by
48
References
2
Citations by year
2026: 42 · 2025: 6
FWCI
36.27
Type
conference paper
Language
en
Open access
Licence
Locations
  • IEEE/CVF International Conference on Computer Vision (ICCV) (publishedVersion)
  • arXiv (Cornell University) (submittedVersion, cc-by)
  • arXiv (Cornell University) (cc-by)
Funders
Ministry of Education, Singapore
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
Keywords
diffusion transformer, image-to-image translation, image generation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review