Research map / 3D vision / 3D generation
Towards High-Fidelity Single-View Holistic Reconstruction of Indoor Scenes
Haolin Liu, Yujian Zheng, Guanying Chen, Shuguang Cui, Xiaoguang Han
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 global instance feature to pixel-aligned feature (ablation) | InstPIFu · 3D-FUTURE objects | Chamfer distance (x10^-3): 17.95 -> 16.42 | Improved |
| add attentional channel filtering to C0 (ablation) | InstPIFu · 3D-FUTURE objects | Chamfer distance (x10^-3): 16.42 -> 15.54 | Improved |
| add spatial-guided mask supervision to C0 (ablation) | InstPIFu · 3D-FUTURE objects | Chamfer distance (x10^-3): 16.42 -> 15.28 | Improved |
| combine channel filtering and mask supervision (ablation) | InstPIFu · 3D-FUTURE objects | F-Score: 58.62 -> 61.32 | Improved |
| predict objects in canonical coordinates instead of camera coordinates (ablation) | InstPIFu · 3D-FUTURE objects | Chamfer distance (x10^-3): 41.64 -> 14.46 | Improved |
| use predicted 2D detections instead of GT boxes at test time (ablation) | InstPIFu · 3D-FUTURE objects | Chamfer distance (x10^-3): 14.46 -> 15.35 | Hurt |
| use instance-aligned implicit background instead of Adabins depth | room background · 3D-FRONT | Chamfer distance: 0.573 -> 0.481 | Improved |
| use InstPIFu instead of MGN (explicit mesh) | object reconstruction · 3D-FUTURE | Chamfer distance (x10^-3): 14.07 -> 14.46 | Hurt |
| use InstPIFu instead of MGN | object reconstruction · 3D-FUTURE | F-Score: 55.64 -> 61.32 | Improved |
| use InstPIFu instead of MGN on prior split | object reconstruction · Pix3D (split of prior work) | Chamfer distance (x10^-3): 6.84 -> 13.60 | Hurt |
| use InstPIFu instead of MGN on unseen shapes | object reconstruction · Pix3D (non-overlapped split) | Chamfer distance (x10^-3): 44.32 -> 24.65 | Improved |
| use InstPIFu instead of baseline for whole scene | holistic scene · 3D-FRONT whole scene | Chamfer distance: 0.290 -> 0.172 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PlaneRCNN — PlaneRCNN: 3D Plane Detection and Reconstruction From a Single Image (CVPR 2019) · better on 1 of 1 shared result
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
- Total3DUnderstanding: Joint Layout, Object Pose and Mesh Reconstruction for Indoor Scenes From a Single Image (CVPR 2020 · similarity 0.90)
- Behind the Veil: Enhanced Indoor 3D Scene Reconstruction with Occluded Surfaces Completion (CVPR 2024 · similarity 0.88)
- Automatic 3D Indoor Scene Modeling From Single Panorama (CVPR 2018 · similarity 0.87)
- Inverse Rendering for Complex Indoor Scenes: Shape, Spatially-Varying Lighting and SVBRDF From a Single Image (CVPR 2020 · similarity 0.87)
- Shading Meets Motion: Self-supervised Indoor 3D Reconstruction Via Simultaneous Shape-from-Shading and Structure-from-Motion (CVPR 2025 · similarity 0.85)
- Sparis: Neural Implicit Surface Reconstruction of Indoor Scenes from Sparse Views (AAAI 2025 · similarity 0.85)
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.
