Research map / 3D vision / Face and shape
Textured 3D Regenerative Morphing with 3D Diffusion Prior
Songlin Yang, Yushi Lan, Honghua Chen, Xingang Pan
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 Attention Fusion on top of basic interpolation (ablation) | 3D diffusion morphing · Textured 3D morphing | smoothness | Improved |
| extend Attention Fusion final time step too far (ablation) | 3D diffusion morphing · Textured 3D morphing | structural plausibility | Hurt |
| add Token Reordering (ablation) | 3D diffusion morphing · Textured 3D morphing | structural plausibility | Improved |
| push Token Reordering time step further (ablation) | 3D diffusion morphing · Textured 3D morphing | generation quality | Hurt |
| add Low-Frequency Enhancement (ablation) | 3D diffusion morphing · Textured 3D morphing | smoothness and plausibility balance | Improved |
| increase low-frequency scale up to 5 (ablation) | 3D diffusion morphing · Textured 3D morphing | surface quality | Improved |
| increase low-frequency scale beyond 5 (ablation) | 3D diffusion morphing · Textured 3D morphing | surface quality | No clear effect |
| use 3D diffusion prior morphing instead of MorphFlow | Ours · Textured 3D morphing | FID: 147.70 -> 6.36 | Improved |
| use 3D diffusion prior morphing instead of Luma video generation | Ours · Textured 3D morphing | FID: 95.49 -> 6.36 | Improved |
| use 3D diffusion prior morphing instead of AID | Ours · Textured 3D morphing | SEP-U: 0.505 -> 0.950 | Improved |
| use 3D diffusion prior morphing instead of DiffMorpher | Ours · Textured 3D morphing | PPL: 5.23 -> 3.02 | Improved |
| use 3D diffusion prior morphing instead of MorphFlow | Ours · Textured 3D morphing | PDV: 0.0001 -> 0.0001 | No clear effect |
Closest papers by meaning
- Wukong's 72 Transformations: High-fidelity Textured 3D Morphing via Flow Models (NeurIPS 2025 · similarity 0.91)
- Sin3DM: Learning a Diffusion Model from a Single 3D Textured Shape (ICLR 2024 · similarity 0.89)
- IMPUS: Image Morphing with Perceptually-Uniform Sampling Using Diffusion Models (ICLR 2024 · similarity 0.88)
- Single Mesh Diffusion Models with Field Latents for Texture Generation (CVPR 2024 · similarity 0.88)
- Kiss3DGen: Repurposing Image Diffusion Models for 3D Asset Generation (CVPR 2025 · similarity 0.88)
- Coherent 3D Scene Diffusion From a Single RGB Image (NeurIPS 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.
