Research map / Point clouds / Shape and surface completion
Primitive-Based Shape Abstraction via Nonparametric Bayesian Inference
Yuwei Wu, Weixiao Liu, Sipu Ruan, Gregory S. Chirikjian
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 merging post-process (ablation) | Ours · ShapeNet (mean) | Chamfer-L1: 0.017 -> 0.021 | Hurt |
| add the merging post-process (ablation) | Ours · ShapeNet (mean) | IoU: 0.547 -> 0.526 | Hurt |
| add the merging post-process (ablation) | Ours · ShapeNet can | number of primitives: 13.6 -> 1.1 | Improved |
| add the merging post-process (ablation) | Ours · ShapeNet chair | number of primitives: 26.9 -> 13.6 | Improved |
| add the merging post-process (ablation) | Ours · ShapeNet table | number of primitives: 25.1 -> 9.5 | Improved |
| replace learned SQs with nonparametric Gibbs-sampling abstraction | Ours vs SQs · ShapeNet (mean) | Chamfer-L1: 0.053 -> 0.021 | Improved |
| replace learned SQs with nonparametric Gibbs-sampling abstraction | Ours vs SQs · ShapeNet (mean) | IoU: 0.242 -> 0.526 | Improved |
| replace K-means clustering with nonparametric Gibbs-sampling abstraction | Ours vs K-means · ShapeNet (mean) | Chamfer-L1: 0.057 -> 0.021 | Improved |
| replace K-means clustering with nonparametric Gibbs-sampling abstraction | Ours vs K-means · ShapeNet (mean) | IoU: 0.410 -> 0.526 | Improved |
| replace learned SQs with nonparametric Gibbs-sampling abstraction | Ours vs SQs · D-FAUST | Chamfer-L1: 0.0473 -> 0.0335 | Improved |
| replace learned SQs with nonparametric Gibbs-sampling abstraction | Ours vs SQs · D-FAUST | IoU: 0.7138 -> 0.7709 | Improved |
| replace learned SQs with nonparametric Gibbs-sampling abstraction | Ours vs SQs · ShapeNet chair | IoU: 0.269 -> 0.557 | Improved |
Closest papers by meaning
- PointGMM: A Neural GMM Network for Point Clouds (CVPR 2020 · similarity 0.88)
- Learning Shape Abstractions by Assembling Volumetric Primitives (CVPR 2017 · similarity 0.87)
- Parameterization-Based Dataset Distillation of 3D Point Clouds through Learnable Shape Morphing (ICLR 2026 · similarity 0.87)
- Accelerated Generative Models for 3D Point Cloud Data (CVPR 2016 · similarity 0.87)
- ShapeOdds: Variational Bayesian Learning of Generative Shape Models (CVPR 2017 · similarity 0.87)
- Learning Shape Primitives via Implicit Convexity Regularization (ICCV 2023 · 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.
