Research map / 3D vision / Depth and stereo
HAMSt3R: Human-Aware Multi-view Stereo 3D Reconstruction
Sara Rojas, Matthieu Armando, Bernard Ghanem, Philippe Weinzaepfel, Vincent Leroy, Grégory Rogez
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 |
|---|---|---|---|
| replace MASt3R encoder with DUNE distilled encoder and add human heads, trained with 50% human data | HAMSt3R · Multi-view depth (5 test sets average) | rel: 3.13 -> 4.26 | Hurt |
| human-aware training on EgoExo4D camera pose | HAMSt3R · EgoExo4D camera pose | AE: 11.70 -> 2.85 (vs MASt3R) | Improved |
| train HAMSt3R on top of DUNE for EgoHumans camera pose | HAMSt3R · EgoHumans camera pose | AE: 3.51 -> 10.24 (DUNE vs HAMSt3R) | Hurt |
| evaluate on large open scenes versus small environments (ablation) | HAMSt3R · EgoHumans | AE: large 16.303; small 2.32 | Mixed |
| joint feed-forward reconstruction with SMPL fitting versus HSfM optimization on EgoExo4D world-coordinates | HAMSt3R · EgoExo4D | W-MPJPE: 0.56 -> 0.51 | Improved |
| HAMSt3R versus HSfM on Procrustes-aligned pose on EgoHumans | HAMSt3R · EgoHumans | PA-MPJPE: 0.05 -> 0.14 | Hurt |
| feed-forward pipeline instead of optimization pipeline (HSfM) | HAMSt3R · Runtime 4-view scene | total runtime: ~118s -> ~32s | Improved |
| human-aware training on RealEstate10K pose regression | HAMSt3R · RealEstate10K | mAA(30): 76.4 -> 77.4 (vs MASt3R) | Improved |
| human-aware training on CO3Dv2 pose regression | HAMSt3R · CO3Dv2 | mAA(30): 81.8 -> 76.3 (vs MASt3R) | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 2 on most shared results.
- DUNE — DUNE: Distilling a Universal Encoder from Heterogeneous 2D and 3D Teachers (CVPR 2025) · better on 2 of 6 shared results
- DUSt3R — DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024) · better on 0 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
- Pow3R: Empowering Unconstrained 3D Reconstruction with Camera and Scene Priors (CVPR 2025 · similarity 0.90)
- DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024 · similarity 0.89)
- MUSt3R: Multi-view Network for Stereo 3D Reconstruction (CVPR 2025 · similarity 0.89)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.89)
- MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Seconds (CVPR 2025 · similarity 0.89)
- ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from Monocular Videos (CVPR 2025 · similarity 0.89)
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.
