mlsanityData to 7 Oct 2026

OmniConsistency: Learning Style-Agnostic Consistency from Paired Stylization Data

arXiv 2505.18445 ↗First posted 24 May 2025cs.CVcited by 32

On the map: Diffusion models → Image generation, text-to-image and editing

Abstract (arXiv)

Diffusion models have advanced image stylization significantly, yet two core challenges persist: (1) maintaining consistent stylization in complex scenes, particularly identity, composition, and fine details, and (2) preventing style degradation in image-to-image pipelines with style LoRAs. GPT-4o's exceptional stylization consistency highlights the performance gap between open-source methods and proprietary models. To bridge this gap, we propose OmniConsistency, a universal consistency plugin leveraging large-scale Diffusion Transformers (DiTs). OmniConsistency contributes: (1) an in-context consistency learning framework trained on aligned image pairs for robust generalization; (2) a two-stage progressive learning strategy decoupling style learning from consistency preservation to mitigate style degradation; and (3) a fully plug-and-play design compatible with arbitrary style LoRAs under the Flux framework. Extensive experiments show that OmniConsistency significantly enhances visual coherence and aesthetic quality, achieving performance comparable to commercial state-of-the-art model GPT-4o.

Key facts. Cited by 32 papers in the map (data through 7 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 43 of its 53 references that are in the map. First posted on arXiv 24 May 2025.

Builds on

43 of 53 references.

Closest ancestors · 12

Same direction, newest first

  1. 2025OminiControl2: Efficient Conditioning for Diffusion Transformers11 Mar
  2. 2025MakeAnything: Harnessing Diffusion Transformers for Multi-Domain Procedural Sequence Generation3 Feb
  3. 2025LayerTracer: Cognitive-Aligned Layered SVG Synthesis via Diffusion Transformer3 Feb
  4. 2024EditWorld: Simulating World Dynamics for Instruction-Following Image Editing23 May
  5. 2024Hunyuan-DiT: A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding14 May
  6. 2024Stable-Makeup: When Real-World Makeup Transfer Meets Diffusion Model12 Mar
  7. 2023IP-Adapter: Text Compatible Image Prompt Adapter for Text-to-Image Diffusion Models13 Aug
  8. 2023T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Models16 Feb
  9. 2023Adding Conditional Control to Text-to-Image Diffusion Models10 Feb
  10. 2022Multi-Concept Customization of Text-to-Image Diffusion8 Dec
  11. 2022InstructPix2Pix: Learning to Follow Image Editing Instructions17 Nov
  12. 2022DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation25 Aug

Other references · 31

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  4. 2023GPT-4 Technical Reportcited by 18,132
  5. 2022Scalable Diffusion Models with Transformerscited by 5,931
  6. 2022Generative Adversarial Networkscited by 5,603
  7. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,290
  8. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,094
  9. 2023Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasetscited by 2,526
  10. 2022Prompt-to-Prompt Image Editing with Cross Attention Controlcited by 2,409
  11. 2021CLIPScore: A Reference-free Evaluation Metric for Image Captioningcited by 2,388
  12. 2021SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equationscited by 1,903
19 more
  1. 2023AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuningcited by 1,418
  2. 2024OminiControl: Minimal and Universal Control for Diffusion Transformercited by 334
  3. 2023DreamSim: Learning New Dimensions of Human Visual Similarity using Synthetic Datacited by 322
  4. 2023ZipLoRA: Any Subject in Any Style by Effectively Merging LoRAscited by 194
  5. 2024InstantStyle: Free Lunch towards Style-Preserving in Text-to-Image Generationcited by 166
  6. 2023SSR-Encoder: Encoding Selective Subject Representation for Subject-Driven Generationcited by 142
  7. 2024Implicit Style-Content Separation using B-LoRAcited by 135
  8. 2025EasyControl: Adding Efficient and Flexible Control for Diffusion Transformercited by 118
  9. 2025Any2AnyTryon: Leveraging Adaptive Position Embeddings for Versatile Virtual Clothing Taskscited by 62
  10. 2021StyleAlign: Analysis and Applications of Aligned StyleGAN Modelscited by 48
  11. 2024FonTS: Text Rendering with Typography and Style Controlscited by 40
  12. 2023StyleAdapter: A Unified Stylized Image Generation Modelcited by 39
  13. 2024ProcessPainter: Learn Painting Process from Sequence Datacited by 37
  14. 2024Stable-Hair: Real-World Hair Transfer via Diffusion Modelcited by 34
  15. 2020Multi-modal Attention for Speech Emotion Recognitioncited by 29
  16. 2025TransAnimate: Taming Layer Diffusion to Generate RGBA Videocited by 24
  17. 2025PhotoDoodle: Learning Artistic Image Editing from Few-Shot Pairwise Datacited by 23
  18. 2025ConsisLoRA: Enhancing Content and Style Consistency for LoRA-based Style Transfercited by 9
  19. 2022CMMD: Cross-Metric Multi-Dimensional Root Cause Analysiscited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 12; 2026: 20.

32 papers.

Sort:

By area:

  1. 2025UniWorld-V1: High-Resolution Semantic Encoders for Unified Visual Understanding and Generationcited by 64
  2. 2025RelationAdapter: Learning and Transferring Visual Relation with Diffusion Transformerscited by 26
  3. 2026FastVMT: Eliminating Redundancy in Video Motion Transfercited by 23
  4. 2025Personalized Vision via Visual In-Context Learningcited by 14
  5. 2025ThinkGen: Generalized Thinking for Visual Generationcited by 12
  6. 2025WorldWander: Bridging Egocentric and Exocentric Worlds in Video Generationcited by 9
  7. 2026StreamingEffect: Real-Time Human-Centric Video Effect Generationcited by 8
  8. 2026VISTA: Triplet-Supervised Video Style Transfer with Diffusion Transformerscited by 8
  9. 2025FeRA: Frequency-Energy Constrained Routing for Effective Diffusion Adaptation Fine-Tuningcited by 7
  10. 2026TeleStyle: Content-Preserving Style Transfer in Images and Videoscited by 5
  11. 2025NumeriKontrol: Adding Numeric Control to Diffusion Transformers for Instruction-based Image Editingcited by 4
  12. 2026OmniHumanoid: Streaming Cross-Embodiment Video Generation with Paired-Free Adaptationcited by 4
20 more
  1. 2025PickStyle: Video-to-Video Style Transfer with Context-Style Adapterscited by 3
  2. 2025CameraMaster: Unified Camera Semantic-Parameter Control for Photography Retouchingcited by 3
  3. 2026ReMoT: Reinforcement Learning with Motion Contrast Tripletscited by 3
  4. 2026CollectionLoRA: Collecting 50 Effects in 1 LoRA via Multi-Teacher On-Policy Distillationcited by 3
  5. 2025MultiEdit: Advancing Instruction-based Image Editing on Diverse and Challenging Taskscited by 2
  6. 2025Unified Long Video Inpainting and Outpainting via Overlapping High-Order Co-Denoisingcited by 1
  7. 2025TokenPure: Watermark Removal through Tokenized Appearance and Structural Guidancecited by 1
  8. 2026ViFeEdit: A Video-Free Tuner of Your Video Diffusion Transformercited by 1
  9. 2026OmniDiT: Extending Diffusion Transformer to Omni-VTON Frameworkcited by 1
  10. 2026EditTransfer++: Toward Faithful and Efficient Visual-Prompt-Guided Image Editingcited by 1
  11. 2026SWEET: Sparse World Modeling with Image Editing for Embodied Task Executioncited by 1
  12. 2026Style-CCL: Content-Preserving Style Transfer via Curriculum Continual Learningcited by 1
  13. 2026Ghosts Beneath Textures: Texture-Relation Cues for Cross-Paradigm AI-Generated Image Detectioncited by 1
  14. 2026UniCSG: Unified High-Fidelity Content-Constrained Style-Driven Generation via Staged Semantic and Frequency Disentanglementcited by 0
  15. 2026Causal Probing for Internal Visual Representations in Multimodal Large Language Modelscited by 0
  16. 2026TeleStyle V2: Beyond Content-Preserving Style Transfer with Self-Distillation and Distribution-Matching-Distillationcited by 0
  17. 2026Staying True to the Origin: Continuous Image Stylization with Smooth Transitionscited by 0
  18. 2026GenPuzzle: Benchmarking Visual Reasoning in Image Generation Modelscited by 0
  19. 2026PAI-Actor: Cinematic Multi-Character Replacement in Dynamic Scenescited by 0
  20. 2026VectorHarness: Recovering Editable, Relation-Preserving Structure from Scientific Graphicscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025OmniStyle: Filtering High Quality Style Transfer Data at Scale
  2. 2025Training-free Stylized Text-to-Image Generation with Fast Inference
  3. 2026CoCoDiff: Correspondence-Consistent Diffusion Model for Fine-grained Style Transfer
  4. 2025Leveraging Diffusion Models for Stylization using Multiple Style Images
  5. 2026SafeStyle: Calibrated Style Residual Injection for Controllable Style-Leakage Trade-off in Diffusion Stylization
  6. 2024ZePo: Zero-Shot Portrait Stylization with Faster Sampling
  7. 2025StyleMe3D: Stylization with Disentangled Priors by Multiple Encoders on 3D Gaussians
  8. 2024Conditional Balance: Improving Multi-Conditioning Trade-Offs in Image Generation