Metraipolis › archive › dc_45fe1108ec7a

GLoD-LVSM: Learning Geometry-Grounded Continuous-LoD Scene Representations for Large View Synthesis Models — arXiv:2610.20254

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.20254
License (as archived)
CC0-1.0
Archived
2026-10-15T21:57:05.933Z
SHA-256 of the archived text
147821f735c81c9711c4c6708649db32898b96237ceeffc26a59278755b4d60a
Stable URL of this page
https://metraipolis.com/documento/dc_45fe1108ec7a

How to cite

GLoD-LVSM: Learning Geometry-Grounded Continuous-LoD Scene Representations for Large View Synthesis Models — arXiv:2610.20254 — arXiv (en), https://arxiv.org/abs/2610.20254 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_45fe1108ec7a (Metraipolis archive dc_45fe1108ec7a, sha256 147821f735c81c9711c4c6708649db32898b96237ceeffc26a59278755b4d60a)
BibTeX
@misc{dc45fe1108ec7a,
  title = {GLoD-LVSM: Learning Geometry-Grounded Continuous-LoD Scene Representations for Large View Synthesis Models — arXiv:2610.20254},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.20254},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_45fe1108ec7a; sha256 147821f735c81c9711c4c6708649db32898b96237ceeffc26a59278755b4d60a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20254. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large view synthesis models (LVSMs) reconstruct scenes from sparse views and render novel views quickly, but they incur unnecessary storage and rendering costs due to token redundancy. We present GLoD-LVSM, a compact, geometry-grounded scene representation for LVSMs with continuous levels of detail. First, instead of learning a fixed set of scene anchor initializations, GLoD-LVSM learns a continuous field that initializes a flexible number of anchors to match the desired budget. A condenser then aggregates scene information into these anchors, with both the field and condenser trained end-to-end. Second, the condenser infers reference-token and anchor geometry to guide compression through geometry-grounded attention, improving rendering quality and enabling frustum culling during decoding. Compared with existing LVSM scene-compression baselines, GLoD-LVSM achieves the best trade-off between rendering quality and scene token count. Compared with the uncompressed base model, GLoD-LVSM reduces the scene-token count by 33.4% and accelerates rendering by $1.23\times$ with no quality loss. Under a relaxed quality-loss budget, the reduction increases to 85.5%, with $3.00\times$ rendering speedup. Further analysis reveals that view-synthesis supervision yields emergent scene depth and self-organized anchors that align with surfaces and adapt their density to local texture complexity. Beyond rendering, we further show that our scene representation also supports scene-semantic understanding, highlighting its potential for broader downstream tasks.

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) ; Graphics (cs.GR).
Pagina del lavoro: https://arxiv.org/abs/2610.20254 — 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 2026-10-15T21:38:29.528Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).