Research map / Point clouds / Shape and surface completion
A Holistic Method for Superquadric Fitting Using Unsupervised Clustering Analysis
Mingyang Zhao, Sipu Ruan, Xiaohong Jia
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 |
|---|---|---|---|
| increase the u log u regularization weight lambda from 0.1 to 3 (ablation) | Superquadric fitting · 100 highly tapered superquadrics | mean and median fitting deviation | Improved |
| replace SuperDec's LM-based parameter optimization with the proposed fitting | SuperDec · ShapeNet test (8,751 models) | L1 error: 1.698 -> 1.346 | Improved |
| replace SuperDec's LM-based parameter optimization with the proposed fitting | SuperDec · ShapeNet test (8,751 models) | L2 error: 0.051 -> 0.035 | Improved |
| replace SuperDec's LM-based parameter optimization with the proposed fitting | SuperDec · ShapeNet test (8,751 models) | number of primitives: 5.8 -> 5.74 | No clear effect |
| use proposed method instead of EMS | EMS · KIT objects | average fitting error: 1.369 -> 1.099 | Improved |
| use proposed method instead of VM fitting | VM · KIT objects | average fitting error: 1.155 -> 1.099 | Improved |
| use proposed method instead of EMS | EMS · KIT objects | fitting time (s): 0.506 -> 0.792 | Hurt |
| use proposed method for cylinder fitting instead of MATLAB built-in | MATLAB built-in cylinder fitting · Cylinder fitting under occlusion | average fitting distance: 2.806 -> 0.781 | Improved |
| use proposed method for sphere fitting instead of MATLAB built-in | MATLAB built-in sphere fitting · Sphere fitting under occlusion | average point-to-sphere distance: 0.0041 -> 0.0072 | Hurt |
| use three axis initializations instead of single PCA initialization (ablation) | Superquadric fitting · Rotated initialization (45 degrees) | fitting success | Improved |
| add coarse segmentation before fitting (ablation) | Superquadric fitting · Scene with floor plane and box (structured noise) | fitting quality | Improved |
Closest papers by meaning
- Robust and Accurate Superquadric Recovery: A Probabilistic Approach (CVPR 2022 · similarity 0.91)
- Revisiting Superquadric Fitting: A Numerically Stable Formulation (TPAMI 2017 · similarity 0.89)
- SuperFlex: Deformable Superquadrics for Point Cloud Decomposition (ECCV 2026 · similarity 0.89)
- Residual Primitive Fitting of 3D Shapes with SuperFrusta (CVPR 2026 · similarity 0.86)
- STAR-Edge: Structure-aware Local Spherical Curve Representation for Thin-walled Edge Extraction from Unstructured Point Clouds (CVPR 2025 · similarity 0.85)
- Correspondence-Free Non-Rigid Point Set Registration Using Unsupervised Clustering Analysis (CVPR 2024 · similarity 0.84)
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.
