Research map / Point clouds / Segmentation and detection
Generic Primitive Detection in Point Clouds Using Novel Minimal Quadric Fits
Tolga Birdal, Benjamin Busam, Nassir Navab, Slobodan Ilić, Peter Franz Sturm
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 |
|---|---|---|---|
| use the minimal 3-point RANSAC with local voting instead of the 4-point variant (ablation) | Quadric detector · Our Dataset | detection accuracy and runtime: order of magnitude speed gain without accuracy loss | Improved |
| use the regularized minimal quadric fit instead of Taubin-42 (42 points) | Quadric fitting · Synthetic quadrics | runtime: runtime advantage of ca. 22% | Improved |
| use the regularized fit (soft scale constraint) instead of the full fit (ablation) | Quadric fitting · Synthetic quadrics with noise 0-5% | fitting error under noise | Improved |
| set weighting parameter w in [0.08, 1.0] instead of too low values (ablation) | Regularized fit (Alg. 2) · Synthetic quadrics, eight noisy points | geometric error: w in [0.08, 1.0] outperforms Taubin; too low w hurts | Improved |
| use the generic type-agnostic detector versus type-specific (sphere) fit (ablation) | Quadric detector · Real datasets | detection rate | Hurt |
| generic detector compared with model-based detectors | Quadric detector · ITODD cylinders (14 scenes) | detection rate (%): 41.9% vs 86% (RANSAC), 72% (PPF3D) | Hurt |
Closest papers by meaning
- A Minimalist Approach to Type-Agnostic Detection of Quadrics in Point Clouds (CVPR 2018 · similarity 0.96)
- Finding Good Configurations of Planar Primitives in Unorganized Point Clouds (CVPR 2022 · similarity 0.90)
- Structure-Aware Surface Reconstruction via Primitive Assembly (ICCV 2023 · similarity 0.87)
- PolyFit: Polygonal Surface Reconstruction From Point Clouds (ICCV 2017 · similarity 0.86)
- H3DNet: 3D Object Detection Using Hybrid Geometric Primitives (ECCV 2020 · similarity 0.86)
- FreePoint: Unsupervised Point Cloud Instance Segmentation (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.
