Research map / Point clouds / Shape and surface completion
PointTriNet: Learned Triangulation of 3D Point Sets
Nicholas Sharp, Maks Ovsjanikov
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 |
|---|---|---|---|
| add neighboring triangles to the classification network input (ablation) | PointTriNet · ShapeNet | watertight (%): 64.6 -> 77.0 | Improved |
| add neighboring triangles to the classification network input (ablation) | PointTriNet · ShapeNet | Chamfer (x100): 0.8310 -> 0.7417 | Improved |
| add the proposal network instead of heuristic neighbor sampling (ablation) | PointTriNet · ShapeNet | watertight (%): 66.4 -> 77.0 | Improved |
| add the proposal network instead of heuristic neighbor sampling (ablation) | PointTriNet · ShapeNet | Chamfer (x100): 0.7748 -> 0.7417 | Improved |
| add the overlap loss (ablation) | PointTriNet · ShapeNet | watertight (%): 76.0 -> 77.0 | Improved |
| add the overlap loss (ablation) | PointTriNet · ShapeNet | Chamfer (x100): 0.7489 -> 0.7417 | Improved |
| replace Scan2Mesh-like dual-graph network with iterative PointNet classifier and proposal | PointTriNet · ShapeNet | watertight (%): 23.3 -> 77.0 | Improved |
| replace Scan2Mesh-like dual-graph network with iterative PointNet classifier and proposal | PointTriNet · ShapeNet | Chamfer (x100): 1.1887 -> 0.7417 | Improved |
| train PointTriNet from scratch on noisy points instead of clean points | PointTriNet · ShapeNet noisy | Chamfer advantage over best classical: 0.7417 vs 0.9195 -> 1.0257 vs 1.2309 | Improved |
| use learned PointTriNet instead of ball pivoting | PointTriNet · ShapeNet | watertight (%): 84.1 -> 77.0 | Hurt |
| use learned PointTriNet instead of ball pivoting | PointTriNet · ShapeNet | Chamfer (x100): 1.3440 -> 0.7417 | Improved |
| remove batch or layer normalization (ablation) | PointTriNet · ShapeNet | result quality | Improved |
Later papers that compared with this method
4 later accepted papers report PointTriNet as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.
- OffsetOPT: Explicit Surface Reconstruction without Normals (CVPR 2025 · better on 1 of 1 shared result)
- DMNet: Delaunay Meshing Network for 3D Shape Representation (ICCV 2023 · better on 2 of 3 shared results)
- CircNet: Meshing 3D Point Clouds with Circumcenter Detection (ICLR 2023 · better on 2 of 2 shared results)
- Learning Delaunay Surface Elements for Mesh Reconstruction (CVPR 2021 · better on 5 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
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (CVPR 2017 · similarity 0.90)
- PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space (NeurIPS 2017 · similarity 0.89)
- CircNet: Meshing 3D Point Clouds with Circumcenter Detection (ICLR 2023 · similarity 0.89)
- PointCNN++: Performant Convolution on Native Points (CVPR 2026 · similarity 0.88)
- PointGrid: A Deep Network for 3D Shape Understanding (CVPR 2018 · similarity 0.88)
- Primal-Dual Mesh Convolutional Neural Networks (NeurIPS 2020 · 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.
