Research map / 3D vision / Face and shape
SG-NN: Sparse Generative Neural Networks for Self-Supervised Scene Completion of RGB-D Scans
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 frame-removal self-supervision with masked loss on unobserved target regions (ablation) | SG-NN · Matterport3D real-world scans | completion of commonly occluded regions | Helped |
| use random crops as self-supervision instead of scan frame removal (ablation) | SG-NN · Matterport3D real-world scans, 2cm | ℓ1 error, entire volume: 0.09 -> 0.13 | Hurt |
| replace TSDF input with point cloud input (ablation) | SG-NN · Matterport3D real-world scans, 2cm | ℓ1 error, entire volume: 0.09 -> 0.15 | Hurt |
| replace TSDF output with occupancy output (ablation) | SG-NN · Matterport3D real-world scans, 2cm | ℓ1 error, entire volume: 0.09 -> 0.13 | Hurt |
| use 2 hierarchy levels (ablation) | SG-NN · Matterport3D real-world scans, 2cm | ℓ1 error, entire volume: 0.09 -> 0.10 | Hurt |
| train on less-complete target scans (ablation) | SG-NN · Matterport3D real-world scans | completion robustness | No clear effect |
Closest papers by meaning
- SPSG: Self-Supervised Photometric Scene Generation From RGB-D Scans (CVPR 2021 · similarity 0.89)
- RGBD Based Dimensional Decomposition Residual Network for 3D Semantic Scene Completion (CVPR 2019 · similarity 0.88)
- RevealNet: Seeing Behind Objects in RGB-D Scans (CVPR 2020 · similarity 0.87)
- Coherent 3D Scene Diffusion From a Single RGB Image (NeurIPS 2024 · similarity 0.86)
- Decompositional Neural Scene Reconstruction with Generative Diffusion Prior (CVPR 2025 · similarity 0.86)
- Boosting Self-Supervision for Single-View Scene Completion via Knowledge Distillation (CVPR 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.