Research map / 3D vision / 3D generation
Graph-to-3D: End-to-End Generation and Manipulation of 3D Scenes Using Scene Graphs
Helisa Dhamo, Fabian Manhardt, Nassir Navab, Federico Tombari
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 |
|---|---|---|---|
| share layout and shape features in one model (ablation) | Graph-to-Box (AtlasNet) · 3DSSG scene generation (constraint accuracy) | total constraint accuracy: 0.89 -> 0.91 | Improved |
| share layout and shape features in one model (ablation) | Graph-to-Box + Graph-to-Shape (AtlasNet) · 3DSSG cycle-consistency with SGPN (object recall top-1) | object recall top-1: 0.51 -> 0.54 | Improved |
| share layout and shape features in one model (ablation) | Graph-to-Box + Graph-to-Shape (AtlasNet) · 3DSSG cycle-consistency with SGPN (triplet recall top-1) | triplet recall top-1: 0.23 -> 0.21 | Hurt |
| add modification network T (ablation) | Graph-to-3D (AtlasNet) · 3DSSG manipulation, relationship change (constraint accuracy) | total constraint accuracy: 0.75 -> 0.80 | Improved |
| add modification network T (ablation) | Graph-to-3D (AtlasNet) · 3DSSG manipulation, node addition (constraint accuracy) | total constraint accuracy: 0.79 -> 0.84 | Improved |
| add modification network T (ablation) | Graph-to-3D (DeepSDF) · 3DSSG manipulation, relationship change (constraint accuracy) | total constraint accuracy: 0.80 -> 0.81 | No clear effect |
| use a GCN for shape generation instead of AtlasNet VAE (ablation) | Graph-to-Box · 3DSSG cycle-consistency with SGPN (object recall top-1) | object recall top-1: 0.41 -> 0.51 | Improved |
| train with canonical object poses (ablation) | Graph-to-Box + Graph-to-Shape · 3DSSG cycle-consistency with SGPN (triplet recall top-1) | triplet recall top-1: 0.05 -> 0.23 | Improved |
| use DeepSDF instead of AtlasNet shape representation (ablation) | Graph-to-3D · 3DSSG scene generation (constraint accuracy) | total constraint accuracy: 0.91 -> 0.92 | No clear effect |
| use Graph-to-3D instead of 3D-SLN | Graph-to-3D (AtlasNet) · 3DSSG scene generation (constraint accuracy) | total constraint accuracy: 0.81 -> 0.91 | Improved |
| use progressive baseline instead of Graph-to-3D for node addition | Progressive · 3DSSG manipulation, node addition (constraint accuracy) | total constraint accuracy: 0.84 -> 0.91 | Hurt |
| use shared model for layout diversity | Graph-to-3D (AtlasNet) · 3DSSG generation diversity (std) | location std: 0.011 -> 0.497 | Improved |
Later papers that compared with this method
3 later accepted papers report Graph-to-3D as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- Planner3D: LLM-enhanced Graph Prior Meets 3D Indoor Scene Explicit Regularization (TPAMI 2025 · better on 9 of 10 shared results)
- EchoScene: Indoor Scene Generation via Information Echo over Scene Graph Diffusion (ECCV 2024 · better on 1 of 1 shared result)
- CommonScenes: Generating Commonsense 3D Indoor Scenes with Scene Graph Diffusion (NeurIPS 2023 · better on 10 of 10 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
- Scene Graph Disentanglement and Composition for Generalizable Complex Image Generation (NeurIPS 2024 · similarity 0.91)
- R3CD: Scene Graph to Image Generation with Relation-Aware Compositional Contrastive Control Diffusion (AAAI 2024 · similarity 0.90)
- GraphDreamer: Compositional 3D Scene Synthesis from Scene Graphs (CVPR 2024 · similarity 0.90)
- SGFormer: Semantic Graph Transformer for Point Cloud-Based 3D Scene Graph Generation (AAAI 2024 · similarity 0.90)
- InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 2024 · similarity 0.89)
- FreeScene: Mixed Graph Diffusion for 3D Scene Synthesis from Free Prompts (CVPR 2025 · similarity 0.89)
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.
