Research map / 3D vision / 3D generation
Atlas: End-to-End 3D Scene Reconstruction from Posed Images
Zak Murez, Tarrence van As, James Bartolozzi, Ayan Sinha, Vijay Badrinarayanan, Andrew Rabinovich
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 |
|---|---|---|---|
| regress TSDF directly from fused image features instead of per-view depth plus TSDF fusion | Atlas · ScanNet 3D geometry | L1: .421 -> .162 | Improved |
| direct 3D regression instead of depth-map fusion | Atlas · ScanNet 2D depth | AbsRel: .087 -> .061 | Improved |
| direct 3D regression instead of depth-map fusion | Atlas · ScanNet 3D geometry | F-score: .558 -> .499 | Hurt |
| direct 3D regression instead of depth-map fusion | Atlas · ScanNet 2D depth | SqRel: .035 -> .042 | Hurt |
| add semantic segmentation head (ablation) | Atlas · ScanNet 2D depth | AbsRel: .061 -> .065 | Hurt |
| add semantic segmentation head (ablation) | Atlas · ScanNet 3D geometry | F-score: .499 -> .520 | Improved |
| add semantic segmentation head (ablation) | Atlas · ScanNet 3D geometry | L1: .162 -> .172 | Hurt |
| reduce number of input frames at inference | Atlas · ScanNet | reconstruction quality: 25 of 784 frames | No clear effect |
| amortize 3D CNN over whole sequence versus per-frame MVS | Atlas · ScanNet inference | per frame time (sec): .071 vs 0.322 | Mixed |
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.
- DPSNet — DPSNet: End-to-end Deep Plane Sweep Stereo (ICLR 2019) · better on 5 of 7 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
- Image-to-Voxel Model Translation for 3D Scene Reconstruction and Segmentation (ECCV 2020 · similarity 0.89)
- VolumeFusion: Deep Depth Fusion for 3D Scene Reconstruction (ICCV 2021 · similarity 0.89)
- 3D Scene Reconstruction from a Single Viewport (ECCV 2020 · similarity 0.89)
- StereoDRNet: Dilated Residual StereoNet (CVPR 2019 · similarity 0.89)
- Large Spatial Model: End-to-end Unposed Images to Semantic 3D (NeurIPS 2024 · similarity 0.87)
- Know Your Neighbors: Improving Single-View Reconstruction via Spatial Vision-Language Reasoning (CVPR 2024 · similarity 0.87)
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.
