mlsanityData to 14 Oct 2026

DuetSVG: Unified Multimodal SVG Generation with Internal Visual Guidance

arXiv 2512.10894 ↗First posted 11 Dec 2025cs.CVcited by 8

On the map: Multimodal LLMs / vision-language models → Chart understanding and anomaly detection

Abstract (arXiv)

Recent vision-language model (VLM)-based approaches have achieved impressive results on SVG generation. However, because they generate only text and lack visual signals during decoding, they often struggle with complex semantics and fail to produce visually appealing or geometrically coherent SVGs. We introduce DuetSVG, a unified multimodal model that jointly generates image tokens and corresponding SVG tokens in an end-to-end manner. DuetSVG is trained on both image and SVG datasets. At inference, we apply a novel test-time scaling strategy that leverages the model's native visual predictions as guidance to improve SVG decoding quality. Extensive experiments show that our method outperforms existing methods, producing visually faithful, semantically aligned, and syntactically clean SVGs across a wide range of applications.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Placed in Chart understanding and anomaly detection, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 45 of its 65 references that are in the map. First posted on arXiv 11 Dec 2025.

Builds on

45 of 65 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025InternSVG: Towards Unified SVG Tasks with Multimodal Large Language Models13 Oct

Other references · 44

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  3. 2023GPT-4 Technical Reportcited by 18,396
  4. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
  5. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  6. 2022Classifier-Free Diffusion Guidancecited by 5,847
  7. 2025Qwen2.5-VL Technical Reportcited by 5,587
  8. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
  9. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,109
  10. 2022DreamFusion: Text-to-3D using 2D Diffusioncited by 3,025
  11. 2025Qwen3-VL Technical Reportcited by 2,793
  12. 2025Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Modelscited by 1,261
32 more
  1. 2023ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillationcited by 1,219
  2. 2025FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Spacecited by 1,006
  3. 2025Emerging Properties in Unified Multimodal Pretrainingcited by 897
  4. 2025Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scalingcited by 812
  5. 2024Autoregressive Model Beats Diffusion: Llama for Scalable Image Generationcited by 777
  6. 2025Show-o2: Improved Native Unified Multimodal Modelscited by 250
  7. 2021CLIPDraw: Exploring Text-to-Drawing Synthesis through Language-Image Encoderscited by 187
  8. 2020DeepSVG: A Hierarchical Generative Network for Vector Graphics Animationcited by 159
  9. 2025Emu3.5: Native Multimodal Models are World Learnerscited by 132
  10. 2022VectorFusion: Text-to-SVG by Abstracting Pixel-Based Diffusion Modelscited by 112
  11. 2023StarVector: Generating Scalable Vector Graphics Code from Images and Textcited by 102
  12. 2025Unified Multimodal Understanding and Generation Models: Advances, Challenges, and Opportunitiescited by 85
  13. 2022Towards Layer-wise Image Vectorizationcited by 79
  14. 2023SVGDreamer: Text Guided SVG Generation with Diffusion Modelcited by 76
  15. 2023IconShop: Text-Guided Vector Icon Synthesis with Autoregressive Transformerscited by 75
  16. 2025OmniSVG: A Unified Scalable Vector Graphics Generation Modelcited by 74
  17. 2024Empowering LLMs to Understand and Generate Complex Vector Graphicscited by 66
  18. 2023Word-As-Image for Semantic Typographycited by 63
  19. 2023DiffSketcher: Text Guided Vector Sketch Synthesis through Latent Diffusion Modelscited by 54
  20. 2023Breathing Life Into Sketches Using Text-to-Video Priorscited by 48
  21. 2025Rendering-Aware Reinforcement Learning for Vector Graphics Generationcited by 39
  22. 2024Chat2SVG: Vector Graphics Generation with Large Language Models and Image Diffusion Modelscited by 38
  23. 2022CLIPVG: Text-Guided Image Manipulation Using Differentiable Vector Graphicscited by 34
  24. 2024Text-to-Vector Generation with Neural Path Representationcited by 33
  25. 2022StyleCLIPDraw: Coupling Content and Style in Text-to-Drawing Translationcited by 28
  26. 2023DeepVecFont-v2: Exploiting Transformers to Synthesize Vector Fonts with Higher Qualitycited by 27
  27. 2024StrokeNUWA: Tokenizing Strokes for Vector Graphic Synthesiscited by 26
  28. 2023VecFusion: Vector Font Generation with Diffusioncited by 20
  29. 2024AniClipart: Clipart Animation with Text-to-Video Priorscited by 17
  30. 2018Vectorization of Line Drawings via PolyVector Fieldscited by 16
  31. 2023ChiroDiff: Modelling chirographic data with Diffusion Modelscited by 14
  32. 2025RoboSVG: A Unified Framework for Interactive SVG Generation with Multi-modal Guidancecited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 8.

8 papers.

  1. 2026Render-in-the-Loop: Vector Graphics Generation via Visual Self-Feedbackcited by 12
  2. 2026VFIG: Vectorizing Complex Figures in SVG with Vision-Language Modelscited by 4
  3. 2026AmodalSVG: Amodal Image Vectorization via Semantic Layer Peelingcited by 3
  4. 2026EffectMaker: Unifying Reasoning and Generation for Customized Visual Effect Creationcited by 2
  5. 2026Bunraku: Turning a Single Illustration into an Editable Live2D Charactercited by 1
  6. 2026A Scalable Vector Graphics Latent Spacecited by 1
  7. 2026Compositional SVG Generation via VLM-Driven Hierarchical Semantic Parsingcited by 0
  8. 2026MSTypography: Multi-character Semantic Typography via Balancing Word Legibility and Object Recognizabilitycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025UniSVG: A Unified Dataset for Vector Graphic Understanding and Generation with Multimodal Large Language Models
  2. 2023Beyond Pixels: Exploring Human-Readable SVG Generation for Simple Images with Vision Language Models
  3. 2026SVG-Score: Human-Aligned Evaluation of Text-to-SVG Generation
  4. 2025SVGauge: Towards Human-Aligned Evaluation for SVG Generation
  5. 2026Evaluating Constrained Iterative Refinement for Scalable Vector Graphics Generation with Off-the-Shelf VLMs
  6. 2025SVGThinker: Instruction-Aligned and Reasoning-Driven Text-to-SVG Generation
  7. 2025SVGenius: Benchmarking LLMs in SVG Understanding, Editing and Generation
  8. 2026IntroSVG: Learning from Rendering Feedback for Text-to-SVG Generation via an Introspective Generator-Critic Framework