AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Supervised Fitting of Geometric Primitives to 3D Point Clouds

Lingxiao Li, Minhyuk Sung, Anastasia Dubrovina, Li Yi, Leonidas J. Guibas

CVPR 2019 · Point clouds › Shape and surface completion · approach: Point cloud upsampling and normals · 205 citations (OpenAlex)

Check 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 changedWhereMeasuredEffect
Predict per-point properties and estimate primitives differentiably instead of directly regressing primitive parametersSPFN vs. DPPN · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 46.99 -> 86.63Improved
Use SPFN-predicted segmentation with Efficient RANSACEfficient RANSAC · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 56.95 -> 74.32Improved
Add predicted primitive types and normals to segmented Efficient RANSACEfficient RANSAC with SPFN segmentation · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 74.32 -> 65.31 -> 61.94Hurt
Add segmentation loss (ablation)SPFN · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 50.04 -> 86.63Improved
Predict point normals and train with normal loss rather than supplying jet-fitting normals from 64k points (ablation)SPFN · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 76.47 -> 86.63Improved
Add fitting residual loss (ablation)SPFN · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 79.81 -> 86.63Improved
Add fitting residual loss (ablation)SPFN · ANSI mechanical component test setSegmentation Mean IoU: 72.70 -> 77.14Improved
Add primitive axis loss (ablation)SPFN · ANSI mechanical component test setPrimitive axis difference (°): 6.27 -> 1.51Improved
Add primitive axis loss (ablation)SPFN · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 80.80 -> 86.63Improved
Use ground-truth rather than predicted primitive types in the fitting residual loss (ablation)SPFN · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 85.25 -> 86.63Improved
Increase test input resolution from 8k to 64k pointsSPFN trained with 8k points · ANSI mechanical component test setPrimitive surface coverage (ε=0.01): 86.63 -> 87.03Improved
Add 10% uniformly sampled outliers during training and testingSPFN · ANSI mechanical component test set with outliersPrimitive surface coverage (ε=0.01): 86.63 -> 82.57Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy