Research map / Point clouds / Shape and surface completion
SuperFlex: Deformable Superquadrics for Point Cloud Decomposition
Gabriel Tavernini, Elisabetta Fedele, Tiago Novello, Leonidas Guibas, Marc Pollefeys, Francis Engelmann
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 |
|---|---|---|---|
| replace SuperDec Chamfer loss with differentiable IoU loss (ablation) | SuperFlex · ShapeNet (13 classes) | IoU: 0.59 -> 0.74 | Improved |
| replace Chamfer loss with IoU loss (ablation) | SuperFlex · ShapeNet (13 classes) | overlap: 4% -> 32% | Hurt |
| add SDF surface loss to IoU loss (ablation) | SuperFlex · ShapeNet (13 classes) | L2 Chamfer: 0.124 -> 0.042 | Improved |
| add SDF surface loss to IoU loss (ablation) | SuperFlex · ShapeNet (13 classes) | overlap: 32% -> 14% | Improved |
| add overlap regularization loss (ablation) | SuperFlex · ShapeNet (13 classes) | overlap: 14% -> 8% | Improved |
| add tapering and bending deformations to rigid superquadrics | SuperFlex · ShapeNet (13 classes) | IoU: 0.70 -> 0.72 | Improved |
| add tapering and bending in test-time refinement (ablation) | SuperFlex · ShapeNet (13 classes) (test-time opt.) | IoU: 0.81 -> 0.85 | Improved |
| add test-time optimization to feed-forward predictions | SuperFlex · ShapeNet (13 classes) | IoU: 0.72 -> 0.87 | Improved |
| add bending direction parameter alpha (ablation) | SuperFlex · ShapeNet (13 classes) (test subset) | IoU: 0.83 -> 0.86 | Improved |
| keep all three bending axes instead of pruning one (ablation) | SuperFlex · ShapeNet (13 classes) | IoU: 0.86 -> 0.87 | Improved |
| keep all three bending axes instead of pruning two (ablation) | SuperFlex · ShapeNet (13 classes) | IoU: 0.84 -> 0.87 | Improved |
| fine-tune with occlusion augmentations | SuperFlex · ASE partial point clouds | IoU: 0.20 -> 0.48 | Improved |
Closest papers by meaning
- SuperDec: 3D Scene Decomposition with Superquadrics Primitives (ICCV 2025 · similarity 0.91)
- A Holistic Method for Superquadric Fitting Using Unsupervised Clustering Analysis (TPAMI 2026 · similarity 0.89)
- Robust and Accurate Superquadric Recovery: A Probabilistic Approach (CVPR 2022 · similarity 0.88)
- Self-supervised Learning of Hybrid Part-aware 3D Representations of 2D Gaussians and Superquadrics (ICCV 2025 · similarity 0.88)
- SymmCompletion: High-Fidelity and High-Consistency Point Cloud Completion with Symmetry Guidance (AAAI 2025 · similarity 0.86)
- SPU-PMD: Self-Supervised Point Cloud Upsampling via Progressive Mesh Deformation (CVPR 2024 · 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.
