Research map / Point clouds / Point cloud classification
P-CORE: Self-Supervised Surface Consistency for Point-Based Neural Editing
Yanshu Zhang, Shichong Peng, Mehran Aghabozorgi, Alireza Moazeni, Ke Li
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 |
|---|---|---|---|
| add self-supervised surface-consistency fine-tuning to PAPR | PAPR · Objaverse zero-shot editing | PSNR: 23.15 -> 28.30 | Improved |
| add self-supervised surface-consistency fine-tuning to PAPR | PAPR · Neural Editor zero-shot editing | PSNR: 24.22 -> 25.31 | Improved |
| use P-CORE instead of Neural Editor (1M points) | Neural Editor · Neural Editor dataset | PSNR: 25.62 -> 25.31 | No clear effect |
| fine-tune on random deformations for realistic motion | PAPR · Mixamo running | PSNR: 24.40 -> 29.95 | Improved |
| fine-tune on random deformations for realistic motion | PAPR · Objaverse hand | PSNR: 28.55 -> 29.66 | Improved |
| remove rays-and-points mode (ablation) | P-CORE · Lego, 2x uniform scaling | hole artifacts (qualitative) | Hurt |
| remove points-only mode (ablation) | P-CORE · Lego, 2x uniform scaling | boundary depth and artifacts (qualitative) | Hurt |
| remove texture consistency loss (ablation) | P-CORE · Lego, 2x uniform scaling | rendering sharpness (qualitative) | Mixed |
| use self-supervised optimized NBS skinning weights instead of RBF weights | P-CORE · Objaverse skinning (6 scenes) | PSNR: 25.32 -> 26.46 | Improved |
| use NBS instead of RBF weights | P-CORE · Objaverse skinning (6 scenes) | Chamfer distance: 0.077 -> 0.077 | No clear effect |
| compare zero-shot P-CORE with supervised PAPR-in-Motion appearance finetune | P-CORE · Objaverse six scenes | PSNR: 33.46 (PAPR-in-Motion) vs 28.84 | Hurt |
| excessive fine-tuning | P-CORE · rendering fidelity | appearance | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PAPR — PAPR: Proximity Attention Point Rendering (NeurIPS 2023) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PointGT: Simultaneous Geometric and Textural Editing for Point-Based Representations (ECCV 2026 · similarity 0.90)
- 4Deform: Neural Surface Deformation for Robust Shape Interpolation (CVPR 2025 · similarity 0.89)
- SPU-PMD: Self-Supervised Point Cloud Upsampling via Progressive Mesh Deformation (CVPR 2024 · similarity 0.87)
- RigPAPR: Rig-Based Animation of Static Neural Point Clouds from a Single-View Video (NeurIPS 2026 · similarity 0.87)
- Diffeomorphic Mesh Deformation via Efficient Optimal Transport for Cortical Surface Reconstruction (ICLR 2024 · similarity 0.87)
- Small Steps and Level Sets: Fitting Neural Surface Models with Point Guidance (CVPR 2024 · 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.
