Metraipolis › archive › dc_a6909ba2d4b5

GeoRefer-Bench: A Benchmark from Referring Pixels to Verifiable Geospatial Reasoning — arXiv:2609.29541

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.29541
License (as archived)
CC0-1.0
Archived
2026-10-08T04:27:35.144Z
SHA-256 of the archived text
81920c374d622241b6d5efc0d51aa2a172ac516a15e3ca3e679d56bef7ef0efa
Stable URL of this page
https://metraipolis.com/documento/dc_a6909ba2d4b5

How to cite

GeoRefer-Bench: A Benchmark from Referring Pixels to Verifiable Geospatial Reasoning — arXiv:2609.29541 — arXiv (en), https://arxiv.org/abs/2609.29541 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a6909ba2d4b5 (Metraipolis archive dc_a6909ba2d4b5, sha256 81920c374d622241b6d5efc0d51aa2a172ac516a15e3ca3e679d56bef7ef0efa)
BibTeX
@misc{dca6909ba2d4b5,
  title = {GeoRefer-Bench: A Benchmark from Referring Pixels to Verifiable Geospatial Reasoning — arXiv:2609.29541},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.29541},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a6909ba2d4b5; sha256 81920c374d622241b6d5efc0d51aa2a172ac516a15e3ca3e679d56bef7ef0efa},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.29541. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Referring segmentation in overhead imagery is inherently relational: a query may ask for the buildings north of the road or the pond closest to a residential area, so the correct referent can contain one object, several objects, or none. Existing benchmarks mainly score mask overlap, which cannot verify whether a model actually resolved the stated spatial relation. We introduce GeoRefer-Bench, a benchmark for verifiable geospatial referring segmentation. Each query is represented by an executable logical form over a metric scene graph, and predictions are evaluated with Exact Query Success (EQS), which is satisfied only when the returned instance set exactly matches the set denoted by the query. GeoRefer-Bench contains 700 whole 2048x2048 UAV scenes (2.94 Gpx) at 12.5 and 25 cm ground sampling distance, 26,217 instances, 142,796 spatial relations, and 20,916 executable queries spanning five reasoning levels. It further includes three paraphrases per query, 24.0% unanswerable queries, 2,477 counterfactual pairs, and five leakage-controlled evaluation splits. An independent audit re-derives object geometry, mask ownership, relation values, query execution, and split provenance, finding zero issues across all 700 scenes. Relation-blind strategies can retain non-trivial mIoU while achieving at most 22.7 EQS overall, showing that overlap alone does not certify relational grounding. Across fifteen current models, the strongest reaches 74.1 EQS but drops from 98.9 at level 1 to 60.5 at level 5, while ten models score below 5 EQS on two-hop queries. GeoRefer-Bench turns geospatial referring segmentation from mask matching into verifiable reference resolution.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.AI).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.29541 — 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 2026-10-08T03:50:58.089Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).