Research map / 3D vision / Face and shape
3D Shape Reconstruction From 2D Images With Disentangled Attribute Flow
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use 3DAttriFlow instead of AtlasNet | ShapeNet 2D-to-3D reconstruction · ShapeNet | L1 Chamfer distance × 10^2: 3.59 -> 3.02 | Helped |
| use 3DAttriFlow instead of SnowflakeNet | MVP point cloud completion · MVP | L2 Chamfer distance × 10^4: 5.86 -> 5.06 | Helped |
| add the semantic sub-pipe (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.16 -> 3.03 | Helped |
| add the geometric sub-pipe (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.41 -> 3.03 | Helped |
| replace the semantic sub-pipe with a simple MLP (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.03 -> 3.12 | Hurt |
| replace the geometric sub-pipe with simple MLPs (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.03 -> 3.08 | Hurt |
| replace both sub-pipes with simple MLPs (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.03 -> 3.21 | Hurt |
| set semantic code dimension to 4 instead of 18 (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.03 -> 3.14 | Hurt |
| set semantic code dimension to 8 instead of 18 (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.03 -> 3.11 | Hurt |
| set semantic code dimension to 32 instead of 18 (ablation) | 3DAttriFlow · ShapeNet 2D-to-3D reconstruction, four-category average | L1-CD × 10^2: 3.03 -> 3.20 | Hurt |
Closest papers by meaning
- D2IM-Net: Learning Detail Disentangled Implicit Fields From Single Images (CVPR 2021 · similarity 0.89)
- MESC-3D:Mining Effective Semantic Cues for 3D Reconstruction from a Single Image (CVPR 2025 · similarity 0.88)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.87)
- $\texttt{NAISR}$: A 3D Neural Additive Model for Interpretable Shape Representation (ICLR 2024 · similarity 0.86)
- Geometry-Guided Semantic Reconstruction for 3D Instance Segmentation (NeurIPS 2026 · similarity 0.86)
- DOCTR: Disentangled Object-Centric Transformer for Point Scene Understanding (AAAI 2024 · similarity 0.85)
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.