Metraipolis › archive › dc_2c68e12892ce

GeoScope: Training-Free Program Execution beyond the Counting Horizon for Ultra-High-Resolution Remote Sensing Question Answering — arXiv:2611.12699

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.12699
License (as archived)
CC0-1.0
Archived
2026-11-10T18:44:14.921Z
SHA-256 of the archived text
c0eeb9c768ef779324bbdbccc3a681d616c9d9836842df60ecec78e359008306
Stable URL of this page
https://metraipolis.com/documento/dc_2c68e12892ce

How to cite

GeoScope: Training-Free Program Execution beyond the Counting Horizon for Ultra-High-Resolution Remote Sensing Question Answering — arXiv:2611.12699 — arXiv (en), https://arxiv.org/abs/2611.12699 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2c68e12892ce (Metraipolis archive dc_2c68e12892ce, sha256 c0eeb9c768ef779324bbdbccc3a681d616c9d9836842df60ecec78e359008306)
BibTeX
@misc{dc2c68e12892ce,
  title = {GeoScope: Training-Free Program Execution beyond the Counting Horizon for Ultra-High-Resolution Remote Sensing Question Answering — arXiv:2611.12699},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.12699},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2c68e12892ce; sha256 c0eeb9c768ef779324bbdbccc3a681d616c9d9836842df60ecec78e359008306},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.12699. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In ultra-high-resolution (UHR) remote sensing images, the objects that questions count in a region or relate by direction often span only a few pixels. Existing methods mainly decide where a vision-language model (VLM) should look and trust its answer in the chosen view. We show that this assumption fails for counting, and measure VLMs as one calibrates a sensor. All eight open VLMs tested fall increasingly short once a view holds more than a certain number of objects, about five for most on DOTA aerial images. This counting horizon narrows as views grow and objects appear smaller. An empirical power-law response model predicts the undercount a split recovers and the spurious counts it adds, and supports the rule of splitting only views beyond the horizon. In GeoScope, a text-only language model parses the question into a typed spatial program that fixes the region and its resolution. The VLM only grounds and counts within its operating range, and directions are computed explicitly from the grounded boxes. To our knowledge, it is the first UHR method to split views by a measured limit of the VLM. Without training, it improves Qwen3-VL-8B by 16.3 points on regional counting and 10.5 on spatial relations in XLRS-Bench, and has the highest reported accuracy, trained methods included, on regional counting and on the whole benchmark by question. On UHR-Micro it improves global and regional counting by 16.8 and 36.2 points, 15.6 of the former from splitting. On MME-RealWorld counting it gains 11.4 points, and we show that an option that is almost never correct yet often chosen distorts comparisons there. 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/2611.12699 — 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 2026-11-10T18:37:37.703Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).