Research map / Point clouds / Shape and surface completion
DMNet: Delaunay Meshing Network for 3D Shape Representation
Chen Zhang, Ganzhangqin Yuan, Wenbing Tao
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 |
|---|---|---|---|
| replace Local Graph Iteration with GCN (ablation) | DMNet · ShapeNet | Chamfer-L1: 0.0230 -> 0.0206 | Improved |
| use DMNet structural feature encoder instead of DeepDT's encoder (ablation) | DMNet · ShapeNet | F-score (0.3%): 0.758 -> 0.804 | Improved |
| use offset distances instead of original vertex coordinates (ablation) | DMNet structural encoder · ShapeNet | Chamfer-L1: 0.0221 -> 0.0206 | Improved |
| add manual structural features (ablation) | DMNet · ShapeNet | Chamfer-L1: 0.0214 -> 0.0206 | Improved |
| add shape constraint loss (ablation) | DMNet · ShapeNet | Chamfer-L1: 0.0210 -> 0.0206 | Improved |
| use DMNet instead of Poisson on sparse 1K points | Surface reconstruction · ShapeNet 1K points | IoU: 0.849 -> 0.912 | Improved |
| use DMNet instead of DSE on FAMOUSTHINGI | Surface reconstruction · FAMOUSTHINGI | Edge Chamfer Distance: 0.726 -> 0.270 | Improved |
| use DMNet instead of DSE on FAMOUSTHINGI | Surface reconstruction · FAMOUSTHINGI | Chamfer Distance: 0.329 -> 0.338 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- PointTriNet — PointTriNet: Learned Triangulation of 3D Point Sets (ECCV 2020) · better on 2 of 3 shared results
- SSRNet+ — SSRNet: Scalable 3D Surface Reconstruction Network (CVPR 2020) · better on 2 of 2 shared results
- DeepDT — DeepDT: Learning Geometry From Delaunay Triangulation for Surface Reconstruction (AAAI 2021) · 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
- DeepDT: Learning Geometry From Delaunay Triangulation for Surface Reconstruction (AAAI 2021 · similarity 0.90)
- GEOMetrics: Exploiting Geometric Structure for Graph-Encoded Objects (ICML 2019 · similarity 0.88)
- Learning Delaunay Surface Elements for Mesh Reconstruction (CVPR 2021 · similarity 0.88)
- Learning Meshing from Delaunay Triangulation for 3D Shape Representation (IJCV 2025 · similarity 0.88)
- Learning Deformable Tetrahedral Meshes for 3D Reconstruction (NeurIPS 2020 · similarity 0.88)
- PolyhedronNet: Representation Learning for Polyhedra with Surface-attributed Graph (ICLR 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.
