Research map / Point clouds / Shape and surface completion
ARO-Net: Learning Implicit Fields From Anchored Radial Observations
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 query-specific anchored radial observations with distance and direction information | ARO-Net · ABC test | HD: 2.78 -> 2.25 | Helped |
| Use query-specific anchored radial observations | ARO-Net · ABC test | IOU: 8.21 -> 8.79 | Helped |
| Use ARO-Net instead of UNDC | 3D reconstruction on ShapeNet chairs · ShapeNet chairs test | LFD: 2.19 -> 1.92 | Helped |
| Use ARO-Net instead of UNDC | 3D reconstruction on ShapeNet airplanes after chair training · ShapeNet airplanes test | LFD: 3.62 -> 3.56 | Helped |
| Reduce the anchor count from 48 to 24 (ablation) | ARO-Net · ABC test | CD: 5.46 -> 5.94 | Hurt |
| Increase the anchor count from 48 to 96 (ablation) | ARO-Net · ABC test | CD: 5.46 -> 5.31 | Helped |
| Replace the attention module and implicit decoder with MLPs (ablation) | ARO-Net · ABC test | CD: 5.46 -> 5.90 | Hurt |
| Use layered Fibonacci anchors instead of uniform anchors (ablation) | ARO-Net · ABC test | CD: 6.29 -> 5.46 | Helped |
| Use layered Fibonacci anchors instead of grid-based anchors (ablation) | ARO-Net · ABC test | CD: 5.63 -> 5.46 | Helped |
| Optimize anchor positions during training rather than keeping them fixed (ablation) | ARO-Net · ABC test | CD: 5.46 -> 5.98 | Hurt |
| Adjust anchor positions for 100 iterations at test time (ablation) | ARO-Net · ABC test | CD: 5.46 -> 5.44 | No clear effect |
| Adjust anchor positions for 200 iterations at test time (ablation) | ARO-Net · ABC test | CD: 5.46 -> 5.75 | Hurt |
Closest papers by meaning
- Learning to Infer Implicit Surfaces without 3D Supervision (NeurIPS 2019 · similarity 0.90)
- Local Deep Implicit Functions for 3D Shape (CVPR 2020 · similarity 0.89)
- RIM-Net: Recursive Implicit Fields for Unsupervised Learning of Hierarchical Shape Structures (CVPR 2022 · similarity 0.89)
- Neural Unsigned Distance Fields for Implicit Function Learning (NeurIPS 2020 · similarity 0.89)
- LISR: Learning Linear 3D Implicit Surface Representation Using Compactly Supported Radial Basis Functions (AAAI 2024 · similarity 0.88)
- A Lightweight UDF Learning Framework for 3D Reconstruction Based on Local Shape Functions (CVPR 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.