Metraipolis › archive › dc_fdfeb36bd796

Auditing GeoFM Evaluation for Field-Extent Segmentation: Label Proxies, Baselines, and When Frozen Features Match Fine-tuning — arXiv:2611.00054

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.00054
License (as archived): CC0-1.0
Archived: 2026-11-03T18:39:17.038Z
SHA-256 of the archived text: e07edb4d717eeef304830808b25c9843b607a1bbae727bb4baaab7936bcf15f3
Stable URL of this page: https://metraipolis.com/documento/dc_fdfeb36bd796

How to cite

Auditing GeoFM Evaluation for Field-Extent Segmentation: Label Proxies, Baselines, and When Frozen Features Match Fine-tuning — arXiv:2611.00054 — arXiv (en), https://arxiv.org/abs/2611.00054 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fdfeb36bd796 (Metraipolis archive dc_fdfeb36bd796, sha256 e07edb4d717eeef304830808b25c9843b607a1bbae727bb4baaab7936bcf15f3)

BibTeX

@misc{dcfdfeb36bd796,
  title = {Auditing GeoFM Evaluation for Field-Extent Segmentation: Label Proxies, Baselines, and When Frozen Features Match Fine-tuning — arXiv:2611.00054},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.00054},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fdfeb36bd796; sha256 e07edb4d717eeef304830808b25c9843b607a1bbae727bb4baaab7936bcf15f3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00054. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Geospatial foundation-model (GeoFM) benchmarks for agricultural field-extent segmentation often rely on cropland land-cover proxies and compare against per-pixel spectral baselines. We measure how these choices affect conclusions using identical single-date Sentinel-2 pixels from six countries. Replacing polygon-derived field labels with ESA WorldCover cropland raises random-forest AUROC from 0.55-0.82 to 0.79-0.96, showing that the proxy creates a more spectrally separable task. On the actual field-extent labels, a from-scratch U-Net reaches 0.89-0.98 AUROC and has higher point AUROC than a Prithvi frozen linear probe in five of six regions. We then compare a frozen backbone and full fine-tuning using the same decoder for Prithvi and TerraMind, with ten matched seeds. At $\varepsilon=0.02$ AUROC, 9 of 12 region$\times$model cells are TOST-equivalent. Kenya is the common failure case: positives cover roughly 0.1% of pixels, and full fine-tuning performs better for both models. Cross-region results are less favorable to frozen models. On AUROC and average precision (AP), the U-Net leads the full 30-transfer matrix, while an exploratory analysis excluding Kenya finds that both frozen decoders outperform their fine-tuned counterparts; no AUROC difference is detected between frozen TerraMind and the U-Net in that subset. For Prithvi, a frozen decoder trains about $110\times$ fewer parameters than full fine-tuning. These results argue for auditing labels and baselines before interpreting GeoFM rankings, and for testing a frozen decoder before assuming that full fine-tuning will help. Our conclusions are limited to single-date, per-pixel field-extent segmentation, not parcel-boundary or instance delineation.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.00054 — 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 2026-11-03T18:12:54.671Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).