Research map / Point clouds / Shape and surface completion
ASHF-Net: Adaptive Sampling and Hierarchical Folding Network for Robust Point Cloud Completion
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 ASHF-Net instead of GRNet | Point cloud completion · ShapeNet Plane, Cabinet, Car, Chair, Lamp, Sofa, Table, Vessel | CD (×10−4): Plane: 1.531 -> 1.398; Cabinet: 3.620 -> 3.490; Car: 2.752 -> 2.322; Chair: 2.945 -> 2.815; Lamp: 2.649 -> 2.519; Sofa: | Helped |
| Use ASHF-Net instead of the strongest baseline for each metric | Point cloud completion · KITTI | FD, MMD, Consistency, Uniformity: FD: 0.434 -> 0.773; MMD: 0.537 -> 0.541; Consistency: 0.313 -> 0.298; Uniformity 0.4%: 0.632 -> 0.602; 0.8%: 0.489 -> 0. | Mixed |
| Add adaptive sampling (AS) (ablation) | ASHF-Net baseline · ShapeNet average | CD (×10−4): 5.276 -> 4.711 | Helped |
| Add gated skip-attention (GSA) (ablation) | ASHF-Net with AS · ShapeNet average | CD (×10−4): 4.711 -> 2.741 | Helped |
| Add multi-resolution supervision (ablation) | ASHF-Net with GSA · ShapeNet average | CD (×10−4): 2.741 -> 2.640 | Helped |
| Add KL regularization (ablation) | ASHF-Net with multi-resolution supervision · ShapeNet average | CD (×10−4): 2.640 -> 2.555 | Helped |
Closest papers by meaning
- Point Cloud Completion by Skip-Attention Network With Hierarchical Folding (CVPR 2020 · similarity 0.92)
- Skeleton-bridged Point Completion: From Global Inference to Local Adjustment (NeurIPS 2020 · similarity 0.90)
- DC-PCN: Point Cloud Completion Network with Dual-Codebook Guided Quantization (AAAI 2025 · similarity 0.89)
- SAMBLE: Shape-Specific Point Cloud Sampling for an Optimal Trade-Off Between Local Detail and Global Uniformity (CVPR 2025 · similarity 0.89)
- PCDreamer: Point Cloud Completion Through Multi-view Diffusion Priors (CVPR 2025 · similarity 0.88)
- PointMAC: Meta-Learned Adaptation for Robust Test-Time Point Cloud Completion (NeurIPS 2025 · similarity 0.88)
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.