Research map / Point clouds / Point cloud classification
PPFNet: Global Context Aware Local Features for Robust 3D Point Matching
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 PPFNet instead of 3DMatch | 3DMatch benchmark before RANSAC · 3DMatch benchmark | recall: 0.5961 -> 0.6231 | Helped |
| Use PPFNet instead of 3DMatch | 3DMatch benchmark after RANSAC · 3DMatch benchmark | recall: 0.67 -> 0.71 | Helped |
| Use PPFNet instead of 3DMatch | 3DMatch benchmark · 3DMatch benchmark | total runtime: 3.21ms -> 2.25ms | Helped |
| Add global context to local features (ablation) | PPFNet · Matching, training set | correct matches at 5% inlier ratio: 48% -> 67% | Helped |
| Add global context to local features (ablation) | PPFNet · Matching, validation set | correct matches at 5% inlier ratio: 46% -> 56% | Helped |
| Add PPF to the input (ablation) | PPFNet · Matching, training set | correct matches at 5% inlier ratio: 65% -> 67% | Helped |
| Add PPF to the input (ablation) | PPFNet · Matching, validation set | correct matches at 5% inlier ratio: 48% -> 56% | Helped |
| Add points and normals to the input (ablation) | PPFNet · Matching, training set | correct matches at 5% inlier ratio: 0% -> 67% | Helped |
| Add points and normals to the input (ablation) | PPFNet · Matching, validation set | correct matches at 5% inlier ratio: 6% -> 56% | Helped |
| Add PPF to the input (ablation) | PPFNet · Rotation matching at 30° | matching performance: 38.3% -> 53.3% | Helped |
| Replace N-tuple loss with contrastive or triplet loss (ablation) | PPFNet · Train and validation correspondence embeddings | separability of matching and non-matching pairs | Hurt |
Closest papers by meaning
- CorrNet3D: Unsupervised End-to-End Learning of Dense Correspondence for 3D Point Clouds (CVPR 2021 · similarity 0.88)
- DGC-GNN: Leveraging Geometry and Color Cues for Visual Descriptor-Free 2D-3D Matching (CVPR 2024 · similarity 0.88)
- The Perfect Match: 3D Point Cloud Matching With Smoothed Densities (CVPR 2019 · similarity 0.88)
- PointCFormer: A Relation-Based Progressive Feature Extraction Network for Point Cloud Completion (AAAI 2025 · similarity 0.87)
- DV-Matcher: Deformation-based Non-rigid Point Cloud Matching Guided by Pre-trained Visual Features (CVPR 2025 · similarity 0.87)
- Multi-View Representation is What You Need for Point-Cloud Pre-Training (ICLR 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.