Research map / Graph learning / Knowledge and temporal graphs
Heterogeneous Graph Learning for Scene Graph Prediction in 3D Point Clouds
Yanni Ma, Hao Liu, Yun Pei, Yulan Guo
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 |
|---|---|---|---|
| treat 3D scene graph as heterogeneous graph with type-specific message passing (replace homogeneous fully-connected KISGP graph) (ablation) | KISGP-based scene graph prediction · 3DSSG PredCls | ngcR@50: 0.791 -> 0.889 | Improved |
| add graph structure updating using HGR predicate scores (ablation) | 3D-HetSGP · 3DSSG PredCls | ngcR@50: 0.811 -> 0.889 | Improved |
| use ground-truth type edges instead of predicted ones (oracle) (ablation) | 3D-HetSGP · 3DSSG PredCls | ngcR@50: 0.889 -> 0.970 | Improved |
| heterogeneous graph without type edges (fully connected) vs homogeneous KISGP graph (ablation) | 3D-HetSGP · 3DSSG PredCls | mR@50: 0.619 -> 0.665 | Improved |
| heterogeneous reasoning vs KISGP on proximity predicates | 3D-HetSGP · 3DSSG PredCls Proximity type | ngcR@50: 61.85 -> 79.07 | Improved |
| heterogeneous reasoning vs KISGP on support predicates | 3D-HetSGP · 3DSSG PredCls Support type | ngcR@50: 94.54 -> 93.22 | Hurt |
| heterogeneous graph reasoning on tail predicate classes | 3D-HetSGP · 3DSSG tail classes | mR@50: 62.52 -> 70.75 | Improved |
| heterogeneous graph reasoning on head predicate classes | 3D-HetSGP · 3DSSG head classes | R@50: 65.00 -> 64.01 | Hurt |
| heterogeneous graph reasoning on SGCls top-k recall | 3D-HetSGP · 3DSSG SGCls | R@50: 0.305 -> 0.298 | No clear effect |
| update type edges from final predicate scores instead of HGSL prediction (ablation) | 3D-HetSGP · 3DSSG type edges, Proximity | type-acc: 0.83 -> 0.91 | Improved |
Closest papers by meaning
- Knowledge-inspired 3D Scene Graph Prediction in Point Cloud (NeurIPS 2021 · similarity 0.90)
- Open3DSG: Open-Vocabulary 3D Scene Graphs from Point Clouds with Queryable Objects and Open-Set Relationships (CVPR 2024 · similarity 0.89)
- SGFormer: Semantic Graph Transformer for Point Cloud-Based 3D Scene Graph Generation (AAAI 2024 · similarity 0.88)
- Object-Centric Representation Learning for Enhanced 3D Semantic Scene Graph Prediction (NeurIPS 2025 · similarity 0.87)
- SG-PGM: Partial Graph Matching Network with Semantic Geometric Fusion for 3D Scene Graph Alignment and Its Downstream Tasks (CVPR 2024 · similarity 0.86)
- DSGG: Dense Relation Transformer for an End-to-end Scene Graph Generation (CVPR 2024 · 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.
