Research map / 3D vision / 3D generation
Indoor Scene Generation From a Collection of Semantic-Segmented Depth Images
Ming-Jia Yang, Yu-Xiao Guo, Bin Zhou, Xin Tong
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 |
|---|---|---|---|
| use a multi-view discriminator instead of a single-view discriminator (ablation) | volumetric GAN · Structured3D-Bedroom | acceptance ratio: decreases from 80.0 to 71.0 | Improved |
| use separate encoders for semantic and depth instead of one unified encoder (ablation) | multi-view discriminator · Structured3D-Bedroom | acceptance ratio: 10.7 acceptance ratio gap (80.0 VS. 69.3) | Improved |
| use one joint multi-view discriminator for depth and semantic instead of separate discriminators (ablation) | multi-view discriminator · Structured3D-Bedroom | acceptance ratio: 16.0 acceptance ratio gap (64.0 vs. 80.0) | Improved |
| use 4 views instead of 2 views in the discriminator (ablation) | multi-view discriminator · Structured3D-Bedroom | acceptance ratio: 70.7 -> 80.0 | Improved |
| use 4 views instead of 6 views in the discriminator (ablation) | multi-view discriminator · Structured3D-Bedroom | acceptance ratio: 72.0 -> 80.0 | Improved |
| use 4 views instead of 8 views in the discriminator (ablation) | multi-view discriminator · Structured3D-Bedroom | acceptance ratio: 70.3 -> 80.0 | Improved |
| increase semantic volume resolution from 32x32x16 to 64x64x32 (ablation) | volumetric GAN · Structured3D-Bedroom | generation quality | No clear effect |
| use extracted object instance volumes for CAD retrieval instead of bounding boxes (ablation) | object retrieval · qualitative scenes | detail of object layouts | Improved |
Closest papers by meaning
- 3D-Aware Indoor Scene Synthesis with Depth Priors (ECCV 2022 · similarity 0.92)
- Fast and Flexible Indoor Scene Synthesis via Deep Convolutional Generative Models (CVPR 2019 · similarity 0.90)
- SSR-2D: Semantic 3D Scene Reconstruction From 2D Images (TPAMI 2024 · similarity 0.88)
- SceneCraft: Layout-Guided 3D Scene Generation (NeurIPS 2024 · similarity 0.88)
- InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 2024 · similarity 0.87)
- Hierarchically-Structured Open-Vocabulary Indoor Scene Synthesis with Pre-trained Large Language Model (AAAI 2025 · 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.
