Metraipolis › archive › dc_48181a150809
Generative Structure from Motion with Native 3D Diffusion — arXiv:2610.35522
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.35522
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-27T08:03:13.670Z
- SHA-256 of the archived text
ad7f74637f6cb39fdd8e9c6ab540c0722cfa1c4c475f73aa66b58e54c8f5afdb- Stable URL of this page
https://metraipolis.com/documento/dc_48181a150809
How to cite
Generative Structure from Motion with Native 3D Diffusion — arXiv:2610.35522 — arXiv (en), https://arxiv.org/abs/2610.35522 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_48181a150809 (Metraipolis archive dc_48181a150809, sha256 ad7f74637f6cb39fdd8e9c6ab540c0722cfa1c4c475f73aa66b58e54c8f5afdb)
BibTeX
@misc{dc48181a150809,
title = {Generative Structure from Motion with Native 3D Diffusion — arXiv:2610.35522},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.35522},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_48181a150809; sha256 ad7f74637f6cb39fdd8e9c6ab540c0722cfa1c4c475f73aa66b58e54c8f5afdb},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.35522. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Recovering an object's 3D structure and camera motion from a few uncalibrated views is a fundamental problem in computer vision. Classical structure from motion methods recover 3D geometry consistent with the input images but cannot recover unseen regions; native generation methods produce plausible 3D objects which are related but not necessarily faithful to the input images. Methods simply concatenating reconstruction and generation allow error to cascade across the two stages, instead of producing a 3D object consistent with both the distribution and the input. We propose GenSfM, which integrates image-based 3D reconstruction with a native 3D diffusion model. In particular, we use a multi-modal diffusion transformer to jointly denoise the shape latent and per-view pose latents under shared self-attention, grounding generated geometry in the observed views while using object-level priors to infer unobserved regions. A compact 3D-native UV-volume tokenization keeps the joint sequence small as views are added, and a pose-aware stage refines geometry and texture by sampling per-voxel image features at the recovered camera poses. On Toys4k and GSO, for $4$ uncalibrated views, GenSfM improves novel-view PSNR by $3.2$-$4.5$ dB and reduces median Chamfer distance by $2.8$-$3.8\times$ over a strong native-3D generation baseline, while recovering input-view poses over $96\%$ Acc@$5^\circ$ without any external pose-estimation backbone, post-hoc optimization, or known intrinsics. Project page: this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.35522 — 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 2026-10-27T07:55:30.995Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).