Research map / 3D vision / 3D generation
RefAny3D: 3D Asset-Referenced Diffusion Models for Image Generation
Hanzhuo Huang, Qingyang Bao, Zekai Gu, Zhongshuo Du, Cheng Lin, Yuan Liu, Sibei Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add shared positional embedding across RGB and point map (ablation) | RefAny3D · 3D asset-referenced generation (qualitative) | geometric consistency | Improved |
| add text-agnostic attention in the point map branch (ablation) | RefAny3D · 3D asset-referenced generation (qualitative) | point map background | Improved |
| add domain-specific LoRA alongside reference LoRA (ablation) | RefAny3D · 3D asset-referenced generation (qualitative) | generation quality | Improved |
| add point map generation branch (ablation) | RefAny3D · 3D asset-referenced generation (qualitative) | 3D consistency | Improved |
| increase conditional views from 4 to 6 to 8 (ablation) | RefAny3D · 3D asset-referenced generation (qualitative) | 3D consistency | Improved |
| use 3D-asset conditioning instead of render-then-edit with Qwen-Image-Edit-2509 | image generation from 3D assets · 3D asset-referenced generation (qualitative) | consistency with 3D asset | Improved |
| use 8-view 3D conditioning over single-view reference baselines | baselines · 3D asset-referenced generation | Overall GPT-eval: 6.367 -> 7.123 | Improved |
Closest papers by meaning
- Kiss3DGen: Repurposing Image Diffusion Models for 3D Asset Generation (CVPR 2025 · similarity 0.91)
- One-2-3-45++: Fast Single Image to 3D Objects with Consistent Multi-View Generation and 3D Diffusion (CVPR 2024 · similarity 0.89)
- World-consistent Video Diffusion with Explicit 3D Modeling (CVPR 2025 · similarity 0.89)
- Diffusion$^2$: Dynamic 3D Content Generation via Score Composition of Video and Multi-view Diffusion Models (ICLR 2025 · similarity 0.89)
- Coherent 3D Scene Diffusion From a Single RGB Image (NeurIPS 2024 · similarity 0.89)
- Generating Images with 3D Annotations Using Diffusion Models (ICLR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
