Research map / 3D vision / Face and shape
Unsupervised 3D Shape Reconstruction by Part Retrieval and Assembly
Read the paperCheck 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 part retrieval and assembly from a user-provided part library instead of Neural Parts | Ours vs. NP · Faucet test | VCD: 0.174 -> 0.134 | Helped |
| use part retrieval and assembly from a user-provided part library instead of Neural Parts | Ours vs. NP · Lamp test | SCD: 0.390 -> 0.303 | Helped |
| use part retrieval and assembly from a user-provided part library instead of whole-shape retrieval | Ours vs. JRD · Chair test | SCD: 0.669 -> 0.539 | Helped |
| add Phase II part shifting (ablation) | Ours · Faucet | VCD: 0.175 -> 0.144 | Helped |
| add Phase II part shifting when Phase III is enabled (ablation) | Ours · Faucet | VCD: 0.171 -> 0.134 | Helped |
| add Phase III part borrowing after Phase II (ablation) | Ours · Faucet | VCD: 0.144 -> 0.134 | Helped |
| reduce the part library from 789 parts to 107 parts (ablation) | Ours · Faucet test | VCD: 0.134 -> 0.225 | Hurt |
| restrict final retrieval to the nearest 5% of parts instead of the full library (ablation) | Ours · Faucet test | VCD: 0.134 -> 0.162 | Hurt |
| use 10 rather than 250 training target shapes for amortized inference (ablation) | Ours · Faucet test | VCD: 0.134 -> 0.150 | Hurt |
| use learned model-based inference instead of nearest-neighbor initialization plus optimization (ablation) | Faucet test · Faucet test | VCD: 0.134 -> 0.185 | Hurt |
Closest papers by meaning
- Learning Unsupervised Hierarchical Part Decomposition of 3D Objects From a Single RGB Image (CVPR 2020 · similarity 0.90)
- Neural Parts: Learning Expressive 3D Shape Abstractions With Invertible Neural Networks (CVPR 2021 · similarity 0.89)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.88)
- X-Part: High Fidelity And Structure Coherent Shape Decomposition And Completion (CVPR 2026 · similarity 0.88)
- Learning to Infer and Execute 3D Shape Programs (ICLR 2019 · similarity 0.87)
- PyTorchGeoNodes: Enabling Differentiable Shape Programs for 3D Shape Reconstruction (CVPR 2025 · similarity 0.87)
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.