Metraipolis › archive › dc_baab9b5cf9ee
Single-Annotation, Multi-Hypothesis Segmentation with Semantic-Spatial Residual Banks — arXiv:2610.41069
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41069
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T08:12:49.329Z
- SHA-256 of the archived text
097e7a2e8fd0e30d270eb25e3cf66f317b0587c1321aae3691e39bf2aaedfe5d- Stable URL of this page
https://metraipolis.com/documento/dc_baab9b5cf9ee
How to cite
Single-Annotation, Multi-Hypothesis Segmentation with Semantic-Spatial Residual Banks — arXiv:2610.41069 — arXiv (en), https://arxiv.org/abs/2610.41069 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_baab9b5cf9ee (Metraipolis archive dc_baab9b5cf9ee, sha256 097e7a2e8fd0e30d270eb25e3cf66f317b0587c1321aae3691e39bf2aaedfe5d)
BibTeX
@misc{dcbaab9b5cf9ee,
title = {Single-Annotation, Multi-Hypothesis Segmentation with Semantic-Spatial Residual Banks — arXiv:2610.41069},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41069},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_baab9b5cf9ee; sha256 097e7a2e8fd0e30d270eb25e3cf66f317b0587c1321aae3691e39bf2aaedfe5d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41069. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Multi-rater segmentation datasets show that a single image can admit several valid masks, yet most training sets expose only one annotation per image. We study single-annotation multi-hypothesis segmentation: each training image exposes one fixed mask, while the model must return a finite bank that covers held-out human annotations at test time. Our experiments reveal a representation tradeoff. Banks learned directly from images recover variations in local boundaries and extent but lose structural hypotheses, whereas frozen semantic features preserve structural grouping but underrepresent spatial variation. We introduce the Semantic-Spatial Residual Bank (SSRB), which freezes a semantic bank trained with balanced assignments and learns a residual from the image for each hypothesis. Across five medical endpoints, SSRB improves coverage of held-out masks at $K=8$ over a Probabilistic U-Net under the same one-mask constraint. On LIDC and LIDC-hard, it also produces substantially more distinct useful masks. On BSDS, it preserves the semantic bank's performance in grouping regions. Its advantage grows as more masks are returned, showing that residuals learned for corresponding heads improve coverage of local annotation variation while preserving how regions are grouped. 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/2610.41069 — 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 2026-11-03T08:02:38.974Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).