Research map / Image generation / Autoregressive and latent generation
MeshMosaic: Scaling Artist Mesh Generation via Local-to-Global Assembly
Rui Xu, Tianyang Xue, Qiujie Dong, Le Wan, Zhe Zhu, Peng Li, Zhiyang Dou, Cheng Lin, Shiqing Xin, Yuan Liu, Wenping Wang, Taku Komura
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 |
|---|---|---|---|
| segment shapes into patches with local quantization and boundary-conditioned autoregressive generation | MeshMosaic vs DeepMesh · Objaverse | F-score: 0.471 -> 0.785 | Improved |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs DeepMesh · Thingi10K | Hausdorff distance: 0.165 -> 0.051 | Improved |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs BPT · ShapeNet | Hausdorff distance: 0.017 -> 0.037 | Hurt |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs BPT · ShapeNet | Chamfer distance L2 x10^3: 0.012 -> 0.019 | Hurt |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs BPT · ShapeNet | Edge Chamfer distance: 0.040 -> 0.052 | Hurt |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs DeepMesh · ShapeNet | Hausdorff distance: 0.037 -> 0.037 | No clear effect |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs DeepMesh · ShapeNet | F-score: 0.791 -> 0.929 | Improved |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs BPT · Objaverse | Edge F-score: 0.164 -> 0.348 | Improved |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs BPT · Thingi10K | F-score: 0.496 -> 0.746 | Improved |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs BPT · User study | Detail Recovery score: 1.084 -> 2.912 | Improved |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs BPT · User study | Neatness score: 1.040 -> 2.780 | Improved |
| segment into patches with local quantization and boundary conditioning | MeshMosaic vs DeepMesh · Complex fighter jet model | triangle count: nearly 30K triangles vs a few hundred to a few thousand | Improved |
Closest papers by meaning
- MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformers (ICLR 2025 · similarity 0.91)
- MeshGPT: Generating Triangle Meshes with Decoder-Only Transformers (CVPR 2024 · similarity 0.90)
- TreeMeshGPT: Artistic Mesh Generation with Autoregressive Tree Sequencing (CVPR 2025 · similarity 0.90)
- ARMesh: Autoregressive Mesh Generation via Next-Level-of-Detail Prediction (NeurIPS 2025 · similarity 0.90)
- MeshArt: Generating Articulated Meshes with Structure-Guided Transformers (CVPR 2025 · similarity 0.89)
- Scaling Mesh Generation via Compressive Tokenization (CVPR 2025 · 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.
