Research map / Point clouds / Shape and surface completion
Surface Reconstruction From Point Clouds by Learning Predictive Context Priors
Check 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 |
|---|---|---|---|
| learn query displacements and conditions jointly to specialize the local context prior (ablation) | Predictive Context Priors · ABC | L2-CD (×1000): 4.21 -> 2.09 | Helped |
| add a predicted query displacement Δq rather than using the global query directly (ablation) | Predictive Context Priors · ABC | L2-CD (×1000): 3.13 -> 2.090 | Helped |
| predict q_l' as a displacement rather than directly (ablation) | Predictive Context Priors · ABC | L2-CD (×1000): 2.78 -> 2.090 | Helped |
| predict the condition f_l' jointly with its corresponding query instead of using autodecoding (ablation) | Predictive Context Priors · ABC | L2-CD (×1000): 4.03 -> 2.090 | Helped |
| train a local context prior and keep it fixed while specializing to the target (ablation) | Predictive Context Priors · ABC | L2-CD (×1000): 4.04 -> 2.09 | Helped |
| keep the local-prior parameters fixed rather than tuning them jointly with the query network (ablation) | Predictive Context Priors · ABC | L2-CD (×1000): 3.60 -> 2.09 | Helped |
| center and scale local regions (ablation) | Local context prior · ABC | L2-CD (×1000): 2.83 -> 2.09 | Helped |
| partition training shapes into 6³ local regions rather than 0³, 4³, or 8³ regions (ablation) | Local context prior · ABC | L2-CD (×1000): 12.06, 4.43, 2.09, 2.58 | Mixed |
| use overlapping local regions (ablation) | Local context prior · ABC | L2-CD (×1000): 2.09 -> 2.09 | No clear effect |
| learn the prior using only target-cloud regions rather than regions across training shapes (ablation) | Local context prior · ABC | L2-CD (×1000): 2.09 -> 2.09 | No clear effect |
| use patch-wise regions instead of part-wise regions (ablation) | Local context prior · ABC | L2-CD (×1000): 2.09 -> 2.9, 2.3, 2.5 | Hurt |
| use the proposed method rather than Neural-Pull | Surface reconstruction · ShapeNet subset (mean) | L2-CD (×100): 0.038 -> 0.0136 | Helped |
Closest papers by meaning
- Inferring Neural Signed Distance Functions by Overfitting on Single Noisy Point Clouds through Finetuning Data-Driven based Priors (NeurIPS 2024 · similarity 0.89)
- A Lightweight UDF Learning Framework for 3D Reconstruction Based on Local Shape Functions (CVPR 2025 · similarity 0.89)
- MultiPull: Detailing Signed Distance Functions by Pulling Multi-Level Queries at Multi-Step (NeurIPS 2024 · similarity 0.89)
- Learning Bijective Surface Parameterization for Inferring Signed Distance Functions from Sparse Point Clouds with Grid Deformation (CVPR 2025 · similarity 0.88)
- OffsetOPT: Explicit Surface Reconstruction without Normals (CVPR 2025 · similarity 0.87)
- Digging into Intrinsic Contextual Information for High-fidelity 3D Point Cloud Completion (AAAI 2025 · similarity 0.87)
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.