Research map / 3D vision / Face and shape
StructEdit: Learning Structural Shape Variations
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 |
|---|---|---|---|
| encode shape deltas directly rather than absolute shapes | StructEdit vs. StructureNet · PartNet geometric-neighborhood reconstruction, average | E_r^geo: 0.911 -> 0.786 | Helped |
| encode shape deltas directly rather than absolute shapes | StructEdit vs. StructureNet · PartNet geometric-neighborhood reconstruction, average | E_r^st: 0.340 -> 0.124 | Helped |
| model edits explicitly in a conditional delta space | StructEdit vs. StructureNet · PartNet edit generation, geometric neighborhoods, average | E_qc^geo: 1.783 -> 1.603 | Helped |
| model edits explicitly in a conditional delta space | StructEdit vs. StructureNet · PartNet edit generation, structural neighborhoods, average | E_qc^st: 0.715 -> 0.579 | Helped |
| encode and decode edits conditioned on the source shape | StructEdit · Synthetic edit transfer | E_t^geo: 0.910 -> 0.604 | Helped |
| encode and decode edits conditioned on the source shape | StructEdit · Synthetic edit transfer | E_t^st: 0.387 -> 0.027 | Helped |
| add skip connections (ablation) | StructEdit · PartNet chairs, geometric neighborhoods | E_r^st: 0.201 -> 0.082 | Helped |
| add skip connections (ablation) | StructEdit · PartNet chairs, structural neighborhoods | E_r^st: 0.364 -> 0.136 | Helped |
| use group normalization (ablation) | StructEdit · PartNet chairs, structural neighborhoods | E_r^st: 0.142 -> 0.136 | Helped |
| use a separate leaf-node classifier (ablation) | StructEdit · PartNet chairs, structural neighborhoods | E_r^st: 0.171 -> 0.136 | Helped |
| encode box parameter deltas rather than modified boxes (ablation) | StructEdit · PartNet chairs, geometric neighborhoods | E_r^geo: 1.737 -> 0.754 | Helped |
Closest papers by meaning
- Learning Part Generation and Assembly for Structure-Aware Shape Synthesis (AAAI 2020 · similarity 0.89)
- RIM-Net: Recursive Implicit Fields for Unsupervised Learning of Hierarchical Shape Structures (CVPR 2022 · similarity 0.89)
- Unsupervised Learning of Intrinsic Structural Representation Points (CVPR 2020 · similarity 0.89)
- ShapeTalk: A Language Dataset and Framework for 3D Shape Edits and Deformations (CVPR 2023 · similarity 0.89)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.88)
- ShapeWalk: Compositional Shape Editing Through Language-Guided Chains (CVPR 2024 · 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.