Metraipolis › archive › dc_ff14a44658e7
Rethinking Asymmetry in Self-Supervised 3D Geometry Learning — arXiv:2610.42092
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.42092
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T08:21:58.489Z
- SHA-256 of the archived text
0fca769d5c24642ca72a5c63da1e70a486d4b2579b5530fea68f4a20514e26d1- Stable URL of this page
https://metraipolis.com/documento/dc_ff14a44658e7
How to cite
Rethinking Asymmetry in Self-Supervised 3D Geometry Learning — arXiv:2610.42092 — arXiv (en), https://arxiv.org/abs/2610.42092 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ff14a44658e7 (Metraipolis archive dc_ff14a44658e7, sha256 0fca769d5c24642ca72a5c63da1e70a486d4b2579b5530fea68f4a20514e26d1)
BibTeX
@misc{dcff14a44658e7,
title = {Rethinking Asymmetry in Self-Supervised 3D Geometry Learning — arXiv:2610.42092},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.42092},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ff14a44658e7; sha256 0fca769d5c24642ca72a5c63da1e70a486d4b2579b5530fea68f4a20514e26d1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.42092. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Self-supervised learning relies on asymmetry between what a model observes and what it must explain. Recent methods learn 3D geometry from unposed images by predicting cameras and scene structure, then rendering held-out views. To avoid trivial solutions, they impose this asymmetry in the input space, restricting how views interact through separate branches or masked attention. We revisit this design with geometric accuracy, rather than rendering quality, as the main objective. When a held-out view lies between context views, the context already constrains the geometry it observes. We therefore impose the asymmetry in the 3D representation space, excluding held-out views only from the reconstructed scene while a shared backbone uses all views for correspondence. This improves depth and pose estimation over input-space designs. We further exploit view density: sparse and dense observations depict the same scene but provide different amounts of correspondence. Aligning sparse-context features with dense-context features supplies a geometric training signal alongside rendering, without a pretrained geometric teacher. We additionally align features with a frozen vision foundation model, which brings further gains. None of these signals requires geometric annotation, and the resulting model improves geometry estimation over prior self-supervised methods. Controlled experiments isolate the contribution of each component, and self-supervised pretraining also improves subsequent supervised training. Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.42092 — 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 2026-11-03T08:06:38.858Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).