Metraipolis › archive › dc_16427ab702ad
Emergent Multi-View Geometry Through Self-Distillation — arXiv:2609.39227
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.39227
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T12:45:03.747Z
- SHA-256 of the archived text
afbf61d584c0d7abf90b91c93b35102628c0267b6ec7af1d9e6a724400f3300c- Stable URL of this page
https://metraipolis.com/documento/dc_16427ab702ad
How to cite
Emergent Multi-View Geometry Through Self-Distillation — arXiv:2609.39227 — arXiv (en), https://arxiv.org/abs/2609.39227 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_16427ab702ad (Metraipolis archive dc_16427ab702ad, sha256 afbf61d584c0d7abf90b91c93b35102628c0267b6ec7af1d9e6a724400f3300c)
BibTeX
@misc{dc16427ab702ad,
title = {Emergent Multi-View Geometry Through Self-Distillation — arXiv:2609.39227},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.39227},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_16427ab702ad; sha256 afbf61d584c0d7abf90b91c93b35102628c0267b6ec7af1d9e6a724400f3300c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39227. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Over a century ago, Henri Poincaré argued that a motionless observer cannot acquire the notion of space. Yet, most visual representation learning methods operate on individual images, while those that leverage multiple views rely on RGB reconstruction, entangling geometry with appearance. We propose Poincar3, a self-supervised method that learns representations from multiple views through self-distillation instead of RGB reconstruction. We combine masked patch and image-level distillation with a teacher that observes additional views, enabling training from scratch without explicit 3D supervision. Poincar3 outperforms both previous single and multi-view self-supervised approaches such as DINOv3, MuM, and Muskie on correspondence estimation, camera pose estimation, and 3D reconstruction. Using a lightweight Poincaré adapter, we also find that our learned features encode camera motion more accurately than existing self-supervised representations. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.AI). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.39227 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T12:14:14.036Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).