Research map / 3D vision / 3D generation
I-Scene: 3D Instance Models are Implicit Generalizable Spatial Learners
Lu Ling, Yunhao Ge, Yichen Sheng, Aniket Bera
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 view-centric (VC) space (ablation) | I-Scene · BlendSwap & Scenethesis | F-Score-S: 79.12 -> 90.79 | Improved |
| add view-centric (VC) space (ablation) | I-Scene · BlendSwap & Scenethesis | CD-S: 0.0351 -> 0.0118 | Improved |
| add view-centric (VC) space (ablation) | I-Scene · BlendSwap & Scenethesis | IoU: 0.7557 -> 0.8222 | Improved |
| add view-centric (VC) space (ablation) | I-Scene · 3D-FRONT | IoU: 0.8598 -> 0.8792 | Improved |
| add non-semantic random scenes (NS) (ablation) | I-Scene · BlendSwap & Scenethesis | F-Score-O: 68.87 -> 72.39 | Improved |
| add non-semantic random scenes (NS) (ablation) | I-Scene · 3D-FRONT | F-Score-O: 81.34 -> 84.28 | Improved |
| add non-semantic random scenes (NS) (ablation) | I-Scene · 3D-FRONT | F-Score-S: 93.77 -> 93.50 | No clear effect |
| train on Rand-25K instead of 3D-FT (ablation) | I-Scene · BlendSwap & Scenethesis | F-Score-S: 90.79 -> 93.60 | Improved |
| train on Rand-25K instead of 3D-FT (ablation) | I-Scene · 3D-FRONT | IoU-B: 0.8792 -> 0.7783 | Hurt |
| scale random scenes from 15K to 25K (ablation) | I-Scene · BlendSwap & Scenethesis | F-Score-S: 92.67 -> 93.60 | Improved |
| mix 3D-FT with Rand-15K instead of 3D-FT only (ablation) | I-Scene · BlendSwap & Scenethesis | F-Score-S: 90.79 -> 94.26 | Improved |
| use I-Scene instead of MIDI | I-Scene vs MIDI · BlendSwap & Scenethesis | CD-S: 0.0212 -> 0.0059 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- MIDI — MIDI: Multi-Instance Diffusion for Single Image to 3D Scene Generation (CVPR 2025) · better on 11 of 11 shared results
- PartCrafter — PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers (NeurIPS 2025) · better on 2 of 3 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
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning (NeurIPS 2025 · similarity 0.88)
- Spatial Understanding from Videos: Structured Prompts Meet Simulation Data (NeurIPS 2025 · similarity 0.88)
- GenSpace: Benchmarking Spatially-Aware Image Generation (NeurIPS 2025 · similarity 0.88)
- From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D (NeurIPS 2025 · similarity 0.88)
- SpaceVista: All-Scale Visual Spatial Reasoning from mm to km (ICML 2026 · similarity 0.87)
- Surprise3D: A Dataset for Spatial Understanding and Reasoning in Complex 3D Scenes (NeurIPS 2025 · 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.
