Research map / 3D vision / Depth and stereo
Learning Scene Coordinate Reconstruction from Unposed Images via Pose Graph Optimization
Tze Ho Elden Tse, Jizong Peng, Angela Yao
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 pose graph construction from scene coordinates instead of vanilla PGO (ablation) | ACE-Zero with PGO · 7-Scenes and Mip-NeRF 360 | PSNR (dB): 12.6 -> 18.5 (7-Scenes); 13.8 -> 21.0 (Mip-NeRF 360) | Improved |
| add epipolar and optical-flow uncertainty weighting (ablation) | ACE-Zero with PGO · 7-Scenes | PSNR (dB): 18.5 -> 21.7 | Improved |
| add uncertainty weighting (ablation) | ACE-Zero with PGO · Mip-NeRF 360 | PSNR (dB): 21.0 -> 24.3 | Improved |
| apply PGO at every ACE-Zero iteration instead of once at the end (ablation) | ACE-Zero with PGO · 7-Scenes | PSNR (dB): 20.5 -> 21.7 | Improved |
| apply PGO at every iteration instead of once at the end (ablation) | ACE-Zero with PGO · Mip-NeRF 360 | PSNR (dB): 23.5 -> 24.3 | Improved |
| fix consistency threshold tau at 0.2m instead of 0.1m (ablation) | pose graph construction · 7-Scenes and Mip-NeRF 360 | PSNR | Hurt |
| add symmetric (bidirectional) consistency check for edges (ablation) | pose graph construction · dense indoor scenes versus sparse settings | PSNR | Mixed |
| add PGO to ACE-Zero | ACE-Zero · Mip-NeRF 360 | PSNR (dB): 22.9 -> 24.3 | Improved |
| add PGO to ACE-Zero | ACE-Zero · Tanks and Temples short, Advanced | PSNR (dB): 14.5 -> 16.8 | Improved |
| add PGO to ACE-Zero | ACE-Zero · 7-Scenes | reconstruction time: 1h -> 30min | Improved |
| add PGO to ACE-Zero | ACE-Zero · 7-Scenes | PSNR (dB): 21.2 -> 21.7 | No clear effect |
Closest papers by meaning
- Learning to Filter Outlier Edges in Global SfM (CVPR 2025 · similarity 0.86)
- LEAP: Liberate Sparse-View 3D Modeling from Camera Poses (ICLR 2024 · similarity 0.86)
- Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis (NeurIPS 2024 · similarity 0.86)
- No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Images (ICLR 2025 · similarity 0.86)
- Learning Structure-from-Motion with Graph Attention Networks (CVPR 2024 · similarity 0.85)
- Light3R-SfM: Towards Feed-forward Structure-from-Motion (CVPR 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.
