Research map / Point clouds / Shape and surface completion
DeepDT: Learning Geometry From Delaunay Triangulation for Surface Reconstruction
Yiming Luo, Zhenxing Mi, 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 |
|---|---|---|---|
| Use DeepDT instead of SSRNet | surface reconstruction on ShapeNet · ShapeNet | Chamfer-L1: 0.24 -> 0.20 | Improved |
| Use DeepDT instead of SSRNet | surface reconstruction on ShapeNet · ShapeNet | NC: 0.967 -> 0.967 | No clear effect |
| Use DeepDT instead of SSRNet | surface reconstruction on DTU · DTU | DC-Mean: 0.30 -> 0.37 | Hurt |
| Use DeepDT instead of SSRNet | surface reconstruction on DTU · DTU | DC-Var.: 0.09 -> 0.15 | Hurt |
| Use DeepDT instead of PSR | surface reconstruction on DTU · DTU | CD-Mean: 1.17 -> 0.68 | Improved |
| Use DeepDT instead of SSRNet-S | ShapeNet-trained reconstruction model evaluated without retraining · Stanford 3D Armadillo | CD Mean: 0.028 -> 0.021 | Improved |
| Use DeepDT instead of SSRNet-S | ShapeNet-trained reconstruction model evaluated without retraining · Stanford 3D Armadillo | CD RMS: 0.131 -> 0.028 | Improved |
| Use DeepDT instead of SSRNet-S | ShapeNet-trained reconstruction model evaluated without retraining · Stanford 3D Bunny | CD Mean: 0.064 -> 0.038 | Improved |
| Use DeepDT instead of SSRNet-S | ShapeNet-trained reconstruction model evaluated without retraining · Stanford 3D Bunny | CD RMS: 0.134 -> 0.095 | Improved |
| Use DeepDT instead of SSRNet-S | ShapeNet-trained reconstruction model evaluated without retraining · Stanford 3D Dragon | CD Mean: 0.053 -> 0.029 | Improved |
| Use DeepDT instead of SSRNet-S | ShapeNet-trained reconstruction model evaluated without retraining · Stanford 3D Dragon | CD RMS: 0.208 -> 0.081 | Improved |
| Use DeepDT instead of SSRNet | DTU scan reconstruction · DTU | time per scan (seconds): 139.652 seconds -> 8.985 seconds | Improved |
Later papers that compared with this method
1 later accepted paper reports DeepDT as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- DMNet: Delaunay Meshing Network for 3D Shape Representation (ICCV 2023 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- SSRNet — SSRNet: Scalable 3D Surface Reconstruction Network (CVPR 2020) · better on 2 of 5 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
- DMNet: Delaunay Meshing Network for 3D Shape Representation (ICCV 2023 · similarity 0.90)
- Learning Delaunay Surface Elements for Mesh Reconstruction (CVPR 2021 · similarity 0.87)
- Deep Geometric Prior for Surface Reconstruction (CVPR 2019 · similarity 0.86)
- Scalable Surface Reconstruction From Point Clouds With Extreme Scale and Density Diversity (CVPR 2017 · similarity 0.86)
- A Survey and Benchmark of Automatic Surface Reconstruction From Point Clouds (TPAMI 2024 · similarity 0.86)
- Surface Reconstruction From Point Clouds: A Survey and a Benchmark (TPAMI 2024 · similarity 0.85)
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.
