Research map / Point clouds / Point cloud classification
Knowledge-inspired 3D Scene Graph Prediction in Point Cloud
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 |
|---|---|---|---|
| add learned meta-embedding knowledge to scene graph prediction (ablation) | Ours · 3DSSG PredCls w/ GC | R@50: 0.592 -> 0.650 | Helped |
| add learned meta-embedding knowledge (ablation) | Ours · 3DSSG SGCls w/ GC | R@50: 0.299 -> 0.300 | No clear effect |
| add learned meta-embedding knowledge (ablation) | Ours · 3DSSG SGCls w/o GC | R@50: 0.338 -> 0.343 | Helped |
| add learned meta-embedding knowledge (ablation) | Ours · 3DSSG PredCls w/o GC | R@50: 0.716 -> 0.784 | Helped |
| add learned meta-embedding knowledge (ablation) | Ours · 3DSSG PredCls | mR@50: 0.410 -> 0.635 | Helped |
| add learned meta-embedding knowledge (ablation) | Ours · 3DSSG SGCls | mR@50: 0.242 -> 0.286 | Helped |
| learn knowledge from class labels via graph auto-encoder instead of from perceptual outputs (vs Schemata) | scene graph prediction · 3DSSG PredCls w/ GC | R@50: 0.582 -> 0.650 | Helped |
| reduce training set of scene graph model to 25% with ME (ablation) | Ours w/ ME · 3DSSG SGCls | mR@50: 0.286 -> 0.219 | Hurt |
| use meta-embedding with 50% of training data (ablation) | Ours w/ ME · 3DSSG SGCls | mR@50: 0.242 -> 0.249 | No clear effect |
| meta-embedding with no perceptual information (0% training) (ablation) | Ours w/ ME · 3DSSG PredCls | mR@50: 0.410 -> 0.428 | Helped |
| meta-embedding intervention on frequent predicate classes | Ours · 3DSSG predicates | R@5 per predicate | Mixed |
Closest papers by meaning
- 3D Spatial Multimodal Knowledge Accumulation for Scene Graph Prediction in Point Cloud (CVPR 2023 · similarity 0.93)
- Object-Centric Representation Learning for Enhanced 3D Semantic Scene Graph Prediction (NeurIPS 2025 · similarity 0.91)
- Open3DSG: Open-Vocabulary 3D Scene Graphs from Point Clouds with Queryable Objects and Open-Set Relationships (CVPR 2024 · similarity 0.90)
- SGFormer: Semantic Graph Transformer for Point Cloud-Based 3D Scene Graph Generation (AAAI 2024 · similarity 0.88)
- Uncovering and Shaping the Latent Representation of 3D Scene Topology in Vision-Language Models (NeurIPS 2026 · similarity 0.87)
- DSGG: Dense Relation Transformer for an End-to-end Scene Graph Generation (CVPR 2024 · 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.