Research map / Graph learning / Graph algorithms and clustering
GEOMetrics: Exploiting Geometric Structure for Graph-Encoded Objects
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 GEOMetrics rather than Pixel2Mesh | ShapeNet single-image reconstruction · ShapeNet 13 classes, mean | F1: 59.72 -> 67.37 | Helped |
| use adaptive face splitting rather than uniform splitting (ablation) | GEOMetrics · ShapeNet chair | F1: 50.33 -> 56.61 | Helped |
| use 0N-GCN rather than standard GCN (ablation) | GEOMetrics · ShapeNet chair | F1: 54.57 -> 56.61 | Helped |
| include the latent mesh loss (ablation) | GEOMetrics · ShapeNet chair | F1: 55.59 -> 56.61 | Helped |
| use the proposed surface sampling losses rather than vertex-to-point loss (ablation) | GEOMetrics · ShapeNet chair | F1: 52.92 -> 56.61 | Helped |
| use GEOMetrics instead of a vertex-matched Pixel2Mesh baseline | ShapeNet chair reconstruction · ShapeNet chair | F1: 38.13 -> 56.61 | Helped |
| use 0N-GCN with max pooling rather than standard GCN (ablation) | Mesh-to-voxel mapping · Mesh-to-voxel mean | MSE: 0.0341 -> 0.0270 | Helped |
| use max pooling rather than NGF aggregation (ablation) | Mesh-to-voxel mapping · Mesh-to-voxel mean | MSE: 0.0318 -> 0.0270 | Helped |
Closest papers by meaning
- Higher-Order Function Networks for Learning Composable 3D Object Representations (ICLR 2020 · similarity 0.89)
- Primal-Dual Mesh Convolutional Neural Networks (NeurIPS 2020 · similarity 0.88)
- MeshGPT: Generating Triangle Meshes with Decoder-Only Transformers (CVPR 2024 · similarity 0.87)
- PolyhedronNet: Representation Learning for Polyhedra with Surface-attributed Graph (ICLR 2025 · similarity 0.86)
- Object-X: Learning to Reconstruct Multi-Modal 3D Object Representations (NeurIPS 2025 · similarity 0.86)
- EdgeMovingNet: Edge-preserving Point Cloud Reconstruction via Joint Geometry Features (CVPR 2025 · similarity 0.86)
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.