Research map / Point clouds / Shape and surface completion
DDM: A Metric for Comparing 3D Shapes Using Directional Distance Fields
Siyu Ren, Junhui Hou, Xiaodong Chen, Hongkai Xiong, Wenping Wang
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 |
|---|---|---|---|
| use DDM instead of CD | template surface fitting · 3DCaricShop | F-0.005: 0.9560 -> 0.9833 | Improved |
| use DDM instead of CD as unsupervised loss | NSFP scene flow · FlyingThings3D | EPE3D (m): 0.0899 -> 0.0657 | Improved |
| use DDM instead of CD as unsupervised loss | SCOOP scene flow · FlyingThings3D | EPE3D (m): 0.0839 -> 0.0732 | Improved |
| use DDM instead of CD in optimization-based scene flow | optimization scene flow · FlyingThings3D | EPE3D (m): 0.1557 -> 0.0843 | Improved |
| refine MAC with DDM | MAC rigid registration · 3DMatch (clean) | SR (%): 82.58 -> 82.77 | Improved |
| refine MAC with DDM on outlier data | MAC rigid registration · 3DMatch (50% outliers) | SR (%): 80.50 -> 78.22 | Hurt |
| add confidence score beta (beta=20 instead of 0) (ablation) | DDM rigid registration · 3DMatch rigid | SR (%): 37.11 -> 82.77 | Improved |
| increase beta from 20 to 30 (ablation) | DDM rigid registration · 3DMatch rigid | SR (%): 82.77 -> 81.78 | Hurt |
| increase reference points M from 1x to 20x (ablation) | DDM rigid registration · 3DMatch rigid | SR (%): 83.20 -> 82.65 | No clear effect |
| increase reference points M from 2e4 to 4e4 (ablation) | DDM non-rigid registration · human motion sequences | RMSE (x1e-2): 0.857 -> 0.612 | Improved |
| increase reference points M from 4e4 to 6e4 (ablation) | DDM non-rigid registration · human motion sequences | RMSE (x1e-2): 0.612 -> 0.580 | Improved |
| increase sigma from 0.05 to 0.1 (ablation) | DDM non-rigid registration · human motion sequences | RMSE (x1e-2): 0.646 -> 0.612 | Improved |
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.
- SCOOP — SCOOP: Self-Supervised Correspondence and Optimization-Based Scene Flow (CVPR 2023) · better on 1 of 1 shared result
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
- Representing 3D Shapes With Probabilistic Directed Distance Fields (CVPR 2022 · similarity 0.87)
- GeoUDF: Surface Reconstruction from 3D Point Clouds via Geometry-guided Distance Representation (ICCV 2023 · similarity 0.86)
- Unsigned Orthogonal Distance Fields: An Accurate Neural Implicit Representation for Diverse 3D Shapes (CVPR 2024 · similarity 0.85)
- Wormhole Loss for Partial Shape Matching (NeurIPS 2024 · similarity 0.84)
- DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling (CVPR 2024 · similarity 0.84)
- Partial-to-Partial Shape Matching with Geometric Consistency (CVPR 2024 · similarity 0.84)
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.
