Research map / 3D vision / 3D generation
Seen2Scene: Completing Realistic 3D Scenes with Visibility-Guided Flow
Quan Meng, Yujin Chen, Lei Li, Matthias Niessner, Angela Dai
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 masked training to VAE and flow matching (ablation) | Seen2Scene · Masked training ablation | Reconstruction L1 (x1e-4): 3.24 -> 2.01 | Improved |
| add masked training (ablation) | Seen2Scene · Masked training ablation | Reconstruction CD (x1e-3): 10.64 -> 9.04 | Improved |
| add masked training (ablation) | Seen2Scene · Masked training ablation | U3D-FPD (x1e2): 20.28 -> 14.56 | Improved |
| add masked training (ablation) | Seen2Scene · Masked training ablation | VLM score: 5.66 -> 5.70 | Improved |
| add real scan data to training (vs synthetic only) (ablation) | Seen2Scene · Generation ablation (ScanNet++, ARKitScenes, 3D-FRONT) | DINOv2-FID (x1e2): 2.90 -> 1.90 | Improved |
| add real scan data to training (vs synthetic only) (ablation) | Seen2Scene · Generation ablation (ScanNet++, ARKitScenes, 3D-FRONT) | U3D-FPD (x1e2): 21.49 -> 14.56 | Improved |
| use CLIP embeddings instead of discrete categories (ablation) | Seen2Scene · Generation ablation (ScanNet++, ARKitScenes, 3D-FRONT) | DINOv2-FID (x1e2): 3.17 -> 1.90 | Improved |
| use CLIP embeddings instead of discrete categories (ablation) | Seen2Scene · Generation ablation (ScanNet++, ARKitScenes, 3D-FRONT) | U3D-FPD (x1e2): 21.46 -> 14.56 | Improved |
| add label synonym augmentation (ablation) | Seen2Scene · Generation ablation (ScanNet++, ARKitScenes, 3D-FRONT) | DINOv2-FID (x1e2): 2.14 -> 1.90 | Improved |
| add label synonym augmentation (ablation) | Seen2Scene · Generation ablation (ScanNet++, ARKitScenes, 3D-FRONT) | U3D-FPD (x1e2): 18.77 -> 14.56 | Improved |
| condition on 3D bounding boxes at completion time (ablation) | Seen2Scene · Scan completion (3D-FRONT only) | Chamfer distance (x1e-2): 1.92 -> 2.05 | Hurt |
| condition on 3D bounding boxes at completion time (ablation) | Seen2Scene · Scan completion (ScanNet++, ARKitScenes, 3D-FRONT) | TMD (x1e-2): 1.69 -> 3.33 | 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.
- SG-NN — SG-NN: Sparse Generative Neural Networks for Self-Supervised Scene Completion of RGB-D Scans (CVPR 2020) · better on 2 of 2 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
- Splatting the Invisible: Geometry and Appearance Scene Completion from Sparse Views (NeurIPS 2026 · similarity 0.89)
- SceneFlowFields++: Multi-frame Matching, Visibility Prediction, and Robust Interpolation for Scene Flow Estimation (IJCV 2019 · similarity 0.88)
- CompNVS: Novel View Synthesis with Scene Completion (ECCV 2022 · similarity 0.87)
- World Tracing: Pixel-Aligned Geometry Beyond the Visible (NeurIPS 2026 · similarity 0.87)
- Learning Temporal 3D Semantic Scene Completion via Optical Flow Guidance (NeurIPS 2025 · similarity 0.86)
- Novel View Synthesis from A Few Glimpses via Test-Time Natural Video Completion (NeurIPS 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.
