Research map / 3D vision / Face and shape
D$^2$CSG: Unsupervised Learning of Compact CSG Trees with Dual Complements and Dropouts
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 |
|---|---|---|---|
| Add complementary primitives (ablation) | CAPRI-Net baseline · ABC | CD: 0.177 -> 0.073 | Helped |
| Add complementary primitives (ablation) | CAPRI-Net baseline · ABC | #P: 66 -> 38 | Helped |
| Add dual branches on top of complementary primitives (ablation) | D2CSG · ABC | CD: 0.073 -> 0.069 | Helped |
| Add dual branches on top of complementary primitives (ablation) | D2CSG · ABC | #P: 38 -> 53 | Hurt |
| Add dropout in final stage on top of CP+DB (ablation) | D2CSG · ABC | #P: 53 -> 29 | Helped |
| Add dropout in final stage on top of CP+DB (ablation) | D2CSG · ABC | CD: 0.069 -> 0.069 | No clear effect |
| Add dropout in final stage on top of CP+DB (ablation) | D2CSG · ABC | NC: 0.936 -> 0.928 | Hurt |
| Replace dual branching by dropout (CP+DO vs CP+DB) (ablation) | D2CSG · ABC | CD: 0.069 -> 0.088 | Hurt |
| Add dropout to CP only (no dual branch) (ablation) | D2CSG · ABC | #P: 38 -> 27 | Helped |
| Add dropout to CP only (no dual branch) (ablation) | D2CSG · ABC | CD: 0.073 -> 0.088 | Hurt |
| Full D2CSG vs CAPRI-Net on mesh input (ablation) | D2CSG · ABC | CD: 0.177 -> 0.069 | Helped |
| Full D2CSG vs CAPRI-Net on mesh input (ablation) | D2CSG · ShapeNet | NC: 0.890 -> 0.887 | Hurt |
Closest papers by meaning
- CAPRI-Net: Learning Compact CAD Shapes With Adaptive Primitive Assembly (CVPR 2022 · similarity 0.91)
- UCSG-NET- Unsupervised Discovering of Constructive Solid Geometry Tree (NeurIPS 2020 · similarity 0.90)
- CSGNet: Neural Shape Parser for Constructive Solid Geometry (CVPR 2018 · similarity 0.87)
- Inferring CAD Modeling Sequences Using Zone Graphs (CVPR 2021 · similarity 0.86)
- Neural Parts: Learning Expressive 3D Shape Abstractions With Invertible Neural Networks (CVPR 2021 · similarity 0.86)
- BSP-Net: Generating Compact Meshes via Binary Space Partitioning (CVPR 2020 · similarity 0.86)
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.