Research map / 3D vision / Face and shape
CraftMesh: High-Fidelity Generative Mesh Manipulation via Poisson Seamless Fusion
James Jincheng Hu, Yuxiao Wu, Youcheng Cai, Ligang Liu
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 Poisson Geometry Blending (ablation) | mesh Boolean baseline · Mesh editing benchmark (100 models x 3 prompts) | CLIPsim: 17.723 -> 20.502 | Improved |
| add Poisson Geometry Blending (ablation) | mesh Boolean baseline · Mesh editing benchmark (100 models x 3 prompts) | NIMA: 5.073 -> 5.290 | Improved |
| add Poisson Texture Harmonization (ablation) | mesh Boolean baseline · Mesh editing benchmark (100 models x 3 prompts) | NIQE: 5.802 -> 5.290 | Improved |
| add Poisson Texture Harmonization (ablation) | mesh Boolean baseline · Mesh editing benchmark (100 models x 3 prompts) | CLIPsim: 17.723 -> 19.399 | Improved |
| combine geometry blending and texture harmonization (ablation) | mesh Boolean baseline · Mesh editing benchmark (100 models x 3 prompts) | CLIPdir: 10.348 -> 18.479 | Improved |
| combine geometry blending and texture harmonization (ablation) | mesh Boolean baseline · Mesh editing benchmark (100 models x 3 prompts) | NIQE: 5.802 -> 4.710 | Improved |
| swap mesh generation backend (MeshyAI, Hunyuan3D 2.5, Trellis) (ablation) | CraftMesh · Mesh editing benchmark (100 models x 3 prompts) | CLIPsim: 20.801 / 19.903 / 19.166 | Mixed |
| use CraftMesh instead of VoxHammer | mesh editing · Mesh editing benchmark (100 models x 3 prompts) | CLIPdir: 10.482 -> 18.479 | Improved |
| remove distribution, gradient, or boundary loss (ablation) | Poisson Texture Harmonization · Mesh editing benchmark (100 models x 3 prompts) | visual quality | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Instant3dit — Instant3dit: Multiview Inpainting for Fast Editing of 3D Objects (CVPR 2025) · better on 4 of 4 shared results
- FocalDreamer — FocalDreamer: Text-Driven 3D Editing via Focal-Fusion Assembly (AAAI 2024) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DreamMesh4D: Video-to-4D Generation with Sparse-Controlled Gaussian-Mesh Hybrid Representation (NeurIPS 2024 · similarity 0.86)
- MeshGen: Generating PBR Textured Mesh with Render-Enhanced Auto-Encoder and Generative Data Augmentation (CVPR 2025 · similarity 0.86)
- Instant3dit: Multiview Inpainting for Fast Editing of 3D Objects (CVPR 2025 · similarity 0.86)
- MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformers (ICLR 2025 · similarity 0.86)
- Scaling Mesh Generation via Compressive Tokenization (CVPR 2025 · similarity 0.86)
- DreamGaussian: Generative Gaussian Splatting for Efficient 3D Content Creation (ICLR 2024 · similarity 0.86)
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.
