Research map / 3D vision / Face and shape
Neural Parts: Learning Expressive 3D Shape Abstractions With Invertible Neural Networks
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 |
|---|---|---|---|
| add the inverse mapping (ablation) | Neural Parts · D-FAUST | IoU: 0.639 -> 0.673 | Helped |
| add the inverse mapping (ablation) | Neural Parts · D-FAUST | Chamfer-L1: 0.119 -> 0.097 | Helped |
| replace the INN with an MLP (ablation) | Neural Parts · D-FAUST | Chamfer-L1: 0.097 -> 0.087 | Mixed |
| add the point projection p_theta (ablation) | Neural Parts · D-FAUST | IoU: 0.638 -> 0.673 | Helped |
| add the point projection p_theta (ablation) | Neural Parts · D-FAUST | Chamfer-L1: 0.106 -> 0.097 | Helped |
| add the occupancy loss (ablation) | Neural Parts · D-FAUST | IoU: 0.642 -> 0.673 | Helped |
| add the reconstruction loss (ablation) | Neural Parts · D-FAUST | Chamfer-L1: 0.150 -> 0.097 | Helped |
| add the normal consistency loss (ablation) | Neural Parts · D-FAUST | IoU: 0.669 -> 0.673 | Helped |
| add the normal consistency loss (ablation) | Neural Parts · D-FAUST | Chamfer-L1: 0.096 -> 0.097 | Mixed |
| add the overlap loss (ablation) | Neural Parts · D-FAUST | IoU: 0.670 -> 0.673 | Helped |
| add the coverage loss (ablation) | Neural Parts · D-FAUST | IoU: 0.668 -> 0.673 | Helped |
| use five expressive primitives instead of ten times as many convex primitives | D-FAUST reconstruction · D-FAUST | IoU: 62% -> 67.3% | Helped |
Closest papers by meaning
- Learning Unsupervised Hierarchical Part Decomposition of 3D Objects From a Single RGB Image (CVPR 2020 · similarity 0.91)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.90)
- Higher-Order Function Networks for Learning Composable 3D Object Representations (ICLR 2020 · similarity 0.90)
- RIM-Net: Recursive Implicit Fields for Unsupervised Learning of Hierarchical Shape Structures (CVPR 2022 · similarity 0.89)
- Unsupervised 3D Shape Reconstruction by Part Retrieval and Assembly (CVPR 2023 · similarity 0.89)
- PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 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.