Research map / 3D vision / Depth and stereo
MAGiSt3R: Multi-Agent Feed-forward 3D Reconstruction from Monocular RGB Videos
ZIREN GONG, Xiaohan Li, Fabio Tosi, Ninghui Xu, Stefano Mattoccia, Jianfei Cai, Matteo Poggi
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 |
|---|---|---|---|
| use MAGMA multi-agent trained merging instead of RANSAC+ICP (ablation) | MAGiSt3R · ReplicaMultiagent multi-agent | accuracy error: 8.89 -> 5.37 | Improved |
| use MAGMA instead of SL(4) optimization (ablation) | MAGiSt3R · ReplicaMultiagent multi-agent | accuracy error: 13.66 -> 5.37 | Improved |
| use MAGMA instead of SLAM3R L2W (ablation) | MAGiSt3R · ReplicaMultiagent multi-agent | accuracy error: 9.71 -> 5.37 | Improved |
| train MAGMA in multi-agent regime instead of single-agent (ablation) | MAGMA · ReplicaMultiagent multi-agent | accuracy error: 6.89 -> 5.37 | Improved |
| train MAGMA in multi-agent regime instead of single-agent (ablation) | MAGMA · ReplicaMultiagent multi-agent | ATE RMSE: 5.35 -> 3.87 | Improved |
| add geometric loss L_geo (ablation) | MAGMA · ReplicaMultiagent multi-agent | accuracy error: 5.88 -> 5.37 | Improved |
| add spatial tokens (ablation) | MAGMA · ReplicaMultiagent multi-agent | accuracy error: 6.04 -> 5.37 | Improved |
| use VGGT backbone instead of SAIL-Recon (ablation) | MAGiSt3R · ReplicaMultiagent multi-agent | accuracy error: 5.97 -> 5.37 | Improved |
| add pose graph optimization (ablation) | MAGiSt3R · ReplicaMultiagent multi-agent | ATE RMSE: 5.65 -> 3.87 | Improved |
| add pose graph optimization (ablation) | MAGiSt3R · ReplicaMultiagent multi-agent | accuracy error: 6.18 -> 5.37 | Improved |
| add MAGMA multi-agent merging versus RICP-merged baselines on AriaMultiagent | MAGiSt3R · AriaMultiagent multi-agent | ATE RMSE: 3.52 vs 6.64 (best RGB baseline SLAM3R) | 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.
- SLAM3R — SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos (CVPR 2025) · 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
- SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos (CVPR 2025 · similarity 0.89)
- MAC-Ego3D: Multi-Agent Gaussian Consensus for Real-Time Collaborative Ego-Motion and Photorealistic 3D Reconstruction (CVPR 2025 · similarity 0.86)
- TrackCraft3R: Repurposing Video Diffusion Transformers for Dense 3D Tracking (NeurIPS 2026 · similarity 0.85)
- Fin3R: Fine-tuning Feed-forward 3D Reconstruction Models via Monocular Knowledge Distillation (NeurIPS 2025 · similarity 0.85)
- EAG3R: Event-Augmented 3D Geometry Estimation for Dynamic and Extreme-Lighting Scenes (NeurIPS 2025 · similarity 0.85)
- MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors (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.
