Metraipolis › archive › dc_7f314c0cdccf

Reconstruction as Spatial Evaluation: Diagnosing Cross-View Spatial Intelligence in Large Vision-Language Models — arXiv:2611.02026

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.02026
License (as archived): CC0-1.0
Archived: 2026-11-03T10:36:21.179Z
SHA-256 of the archived text: aea6e7d2a43f1269bcedfb32358eabf9fa8b82a60c23cfe1dba3bf16beca59da
Stable URL of this page: https://metraipolis.com/documento/dc_7f314c0cdccf

How to cite

Reconstruction as Spatial Evaluation: Diagnosing Cross-View Spatial Intelligence in Large Vision-Language Models — arXiv:2611.02026 — arXiv (en), https://arxiv.org/abs/2611.02026 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7f314c0cdccf (Metraipolis archive dc_7f314c0cdccf, sha256 aea6e7d2a43f1269bcedfb32358eabf9fa8b82a60c23cfe1dba3bf16beca59da)

BibTeX

@misc{dc7f314c0cdccf,
  title = {Reconstruction as Spatial Evaluation: Diagnosing Cross-View Spatial Intelligence in Large Vision-Language Models — arXiv:2611.02026},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02026},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7f314c0cdccf; sha256 aea6e7d2a43f1269bcedfb32358eabf9fa8b82a60c23cfe1dba3bf16beca59da},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02026. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

The spatial intelligence of large vision-language models (LVLMs) requires the ability to combine local visual observations into a geometrically consistent representation of the physical world. Many static benchmarks assess isolated spatial judgments, such as relative direction or distance, while interactive embodied evaluations combine perception with exploration, planning, control, and simulator-specific execution. To bridge this gap, we introduce RECONSPATIAL, a benchmark centered on multi-view 3D scene reconstruction across indoor, tabletop, and outdoor environments. Given real multi-view images and target-object queries, LVLMs are required to jointly predict oriented 3D bounding boxes in a shared, camera-anchored metric coordinate frame. Reconstructions are evaluated along two complementary axes: Structure for relative layout and Metric for physical scale and camera-anchored placement. Nine atomic spatial QA tasks, grouped into Geometric Perception and Spatial Relations, probe spatial capabilities involved in joint reconstruction. Offline navigation extends evaluation to path planning. Evaluation of 18 frontier LVLMs reveals a consistent gap between relative scene structure and metric reconstruction; even GPT-6-Astra scores 81.4% on Structure but only 58.7% on Metric. Tracing this gap, diagnostic analyses link atomic spatial capabilities to reconstruction quality and show that relational consistency alone does not ensure metric accuracy. Experiments with 3D map injection further separate scene construction from path planning, revealing how these geometric errors limit navigation success. Finally, spatial fine-tuning yields broad transfer across both spatial and general vision-language benchmarks.

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/2611.02026 — 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 2026-11-03T10:29:48.834Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).