Research map / 3D vision / 3D generation
Neural Scene Decoration from a Single Photograph
Hong-Wing Pang, Yingshu Chen, Phuoc-Hieu Le, Binh-Son Hua, Thanh Nguyen, Sai-Kit Yeung
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 object discriminator D_obj (ablation) | Ours (D_adv) · Structured3D bedroom, point label | FID: 21.069 -> 15.108 | Improved |
| add object discriminator D_obj (ablation) | Ours (D_adv) · Structured3D bedroom, box label | FID: 22.341 -> 20.596 | Improved |
| add object discriminator D_obj (ablation) | Ours (D_adv) · Structured3D living room, box label | FID: 25.051 -> 18.478 | Improved |
| add object discriminator D_obj (ablation) | Ours (D_adv) · Structured3D living room, point label | KID x10^3: 14.562 -> 9.421 | Improved |
| condition discriminator on background X (ablation) | Ours · Structured3D living room, point label | FID: 17.986 -> 21.653 | Hurt |
| condition discriminator on background X (ablation) | Ours · Structured3D bedroom, box label | FID: 20.596 -> 20.511 | No clear effect |
| use ground-truth object sizes instead of default sizes | Ours (point label) · Structured3D bedroom | FID: 17.489 -> 15.108 | Improved |
| use class median rather than class mean for default object size (ablation) | Ours (point label) · Structured3D living room | FID: 27.365 -> 21.019 | Improved |
| set size multiplier m=2.5 instead of m=4.0 (ablation) | Ours (point label) · Structured3D bedroom | FID: 24.045 -> 15.108 | Improved |
| set size multiplier m=2.5 instead of m=1.0 (ablation) | Ours (point label) · Structured3D bedroom | FID: 20.201 -> 15.108 | Improved |
| use Ours instead of SPADE baseline | SPADE · Structured3D bedroom, point label | FID: 20.345 -> 15.108 | Improved |
| use Ours instead of BachGAN | BachGAN · Structured3D bedroom, box label | KID x10^3: 10.054 vs 11.609 | Hurt |
Closest papers by meaning
- S-INF: Towards Realistic Indoor Scene Synthesis via Scene Implicit Neural Field (AAAI 2025 · similarity 0.89)
- SceneCraft: Layout-Guided 3D Scene Generation (NeurIPS 2024 · similarity 0.89)
- Hierarchically-Structured Open-Vocabulary Indoor Scene Synthesis with Pre-trained Large Language Model (AAAI 2025 · similarity 0.87)
- InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 2024 · similarity 0.87)
- DeBaRA: Denoising-Based 3D Room Arrangement Generation (NeurIPS 2024 · similarity 0.86)
- Architect: Generating Vivid and Interactive 3D Scenes with Hierarchical 2D Inpainting (NeurIPS 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.
