Metraipolis › archive › dc_b72ef7152f22

What Survives Label Downsampling? Exact Region Survival for Coarse Segmentation Targets — arXiv:2610.39880

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.39880
License (as archived)
CC0-1.0
Archived
2026-11-01T02:42:12.330Z
SHA-256 of the archived text
42d657cda9389c1cf616436f761028bfe007dd2571c0b01b121a51671a97924b
Stable URL of this page
https://metraipolis.com/documento/dc_b72ef7152f22

How to cite

What Survives Label Downsampling? Exact Region Survival for Coarse Segmentation Targets — arXiv:2610.39880 — arXiv (en), https://arxiv.org/abs/2610.39880 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b72ef7152f22 (Metraipolis archive dc_b72ef7152f22, sha256 42d657cda9389c1cf616436f761028bfe007dd2571c0b01b121a51671a97924b)
BibTeX
@misc{dcb72ef7152f22,
  title = {What Survives Label Downsampling? Exact Region Survival for Coarse Segmentation Targets — arXiv:2610.39880},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.39880},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b72ef7152f22; sha256 42d657cda9389c1cf616436f761028bfe007dd2571c0b01b121a51671a97924b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39880. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Segmentation networks that predict at an output stride are trained against label maps downsampled by strided sampling, nearest-neighbor resampling or majority pooling. These methods label each coarse grid cell on its own, so a region that wins no cell is absent from the target and receives no positive label. We present MatchDown, which builds the coarse target from a minimum-cost matching between regions and the grid cells they occupy. It keeps as many regions as any one-region-per-cell target can, certifies how many losses are unavoidable, and with priority tiers (MatchDown-P) keeps every evaluated region (a region that a metric scores) of at least one cell's area. On COCO, Cityscapes and BBBC038 at factors 2 to 16, MatchDown-P loses exactly the minimum number of regions, whereas the downsampling used by StarDist, nnU-Net and connectomics pyramids loses four to five times as many on COCO at factor 8. Keeping a region in the target and reconstructing its shape are different problems: we show that no region lost by majority pooling can be reconstructed above IoU 0.5 by upsampling any coarse target, yet on BBBC038 a star-convex head trained on MatchDown targets segments 35% of the nuclei that majority pooling loses above IoU 0.5, against 15% when trained on majority targets and 24% on StarDist's strided target. On two nucleus datasets, with thresholds cross-validated for count error, this head has instance accuracy at IoU 0.5 (Acc50) 2.6 to 4.2 points above the strided target and at most 2.8 points below cell-center sampling, and a lower count error than the strided and cell-center targets on BBBC039. On BBBC039, an affinity head, which outputs one label per cell, lowers count error relative to majority targets by 9.3 nuclei per image at 7.2 points lower Acc50. MatchDown runs once, before training, in place of the label-downsampling step.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.39880 — 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 2026-11-01T02:16:43.955Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).