Metraipolis › archive › dc_cc668433f83f

RDGSplat: Render-Dedicated Geometry for Novel View Synthesis — arXiv:2610.09173

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.09173
License (as archived)
CC0-1.0
Archived
2026-10-08T12:46:57.980Z
SHA-256 of the archived text
b545ae040823eb4a546ffd80033a8d4ffb7d259e8d01230e32fa3b698ca5cacb
Stable URL of this page
https://metraipolis.com/documento/dc_cc668433f83f

How to cite

RDGSplat: Render-Dedicated Geometry for Novel View Synthesis — arXiv:2610.09173 — arXiv (en), https://arxiv.org/abs/2610.09173 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cc668433f83f (Metraipolis archive dc_cc668433f83f, sha256 b545ae040823eb4a546ffd80033a8d4ffb7d259e8d01230e32fa3b698ca5cacb)
BibTeX
@misc{dccc668433f83f,
  title = {RDGSplat: Render-Dedicated Geometry for Novel View Synthesis — arXiv:2610.09173},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.09173},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cc668433f83f; sha256 b545ae040823eb4a546ffd80033a8d4ffb7d259e8d01230e32fa3b698ca5cacb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09173. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

3D foundation models enable efficient novel view synthesis by carrying a Gaussian head on the representation they already use for reconstruction. However, the views they render fall short of the geometry they recover, because that geometry is estimated under a metric objective and never scored on how it renders. Recent methods alleviate this by updating the backbone weights, but they thereby discard the metric predictions the model was built for and must be repeated for every new backbone. To this end, we propose RDGSplat, a framework that decodes a second geometry dedicated to rendering from a frozen 3D foundation model, leaving its metric predictions intact. In particular, we devise Render-Dedicated Geometry Decoding, which duplicates the pretrained decoders and optimizes the duplicates under photometric supervision alone. Then, a Target-Pose Conditioned Adapter is introduced to reformulate the representation those decoders read, conditioned on the target camera pose rather than the target image. Extensive experiments show that RDGSplat improves novel view synthesis across three feed-forward backbones on four benchmarks, with every pretrained weight frozen. On RE10K, it raises WM2.0 from 20.918 to 24.266\,dB while training 205.5\,M added parameters against a frozen 1.4\,B backbone, and the depth and pose the same model predicts are unchanged.

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/2610.09173 — 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 2026-10-08T12:18:15.313Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).