Metraipolis › archive › dc_0293b0bd803b

PAGER: Partial-to-global Alignment via Geometric and Relational Distillation — arXiv:2610.01589

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.01589
License (as archived)
CC0-1.0
Archived
2026-10-06T19:13:55.524Z
SHA-256 of the archived text
ff843d494b7e265db58a8b0af1560febaa9fc570eec96349f327821fe6298e05
Stable URL of this page
https://metraipolis.com/documento/dc_0293b0bd803b

How to cite

PAGER: Partial-to-global Alignment via Geometric and Relational Distillation — arXiv:2610.01589 — arXiv (en), https://arxiv.org/abs/2610.01589 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0293b0bd803b (Metraipolis archive dc_0293b0bd803b, sha256 ff843d494b7e265db58a8b0af1560febaa9fc570eec96349f327821fe6298e05)
BibTeX
@misc{dc0293b0bd803b,
  title = {PAGER: Partial-to-global Alignment via Geometric and Relational Distillation — arXiv:2610.01589},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.01589},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0293b0bd803b; sha256 ff843d494b7e265db58a8b0af1560febaa9fc570eec96349f327821fe6298e05},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.01589. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Pretrained 3D encoders are typically developed on globally reconstructed scenes expressed in a consistent world coordinate frame, whereas embodied systems must reason from partial, viewpoint-dependent observations in camera coordinates. We show that this shift from globally learned 3D feature spaces to realistic partial observations exposes a severe representation mismatch, which we find consistently across representative state-of-the-art encoders, including Sonata and Concerto. A frozen Sonata encoder with a global linear probe achieves 72.47 mIoU on full ScanNet scenes, but 2.57 mIoU on single-frame camera-coordinate inputs. Training-free gravity alignment recovers performance to 41.64 mIoU, showing that coordinate-frame mismatch is a dominant source of degradation but cannot be fully resolved through canonicalization alone. We introduce PAGER, a label-free adaptation method that aligns partial-view features with a frozen global 3D semantic space using only paired partial/global geometry. It learns lightweight adaptation modules while keeping the pretrained encoder and global segmentation probe frozen. Matched-point feature alignment anchors partial features to their global counterparts, while relational supervision preserves their similarity structure with respect to the global representation. Global geometry provides supervision only during training. Inference operates directly on the partial observation. Without partial-view labels, PAGER outperforms label-supervised PEFT on both Sonata and Concerto, and in zero-shot ScanNet$\rightarrow$ScanNet++ transfer surpasses fully fine-tuned Sonata ($53.93$ vs.\ $48.09$ mIoU), suggesting that preserving the frozen global representation can improve cross-dataset transfer.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-10-02 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.01589 — 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 2026-10-06T17:37:36.808Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).