Metraipolis › archive › dc_e62f84e1e8ff
How Many, but How Sure? Class Competition as a Confidence Proxy in Multi-Class Density-Based Counting — arXiv:2610.23611
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.23611
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T09:10:20.755Z
- SHA-256 of the archived text
4e253c3663085f3239aada64dab21cdb81c03115d54a78d9ed57d9025d051cbd- Stable URL of this page
https://metraipolis.com/documento/dc_e62f84e1e8ff
How to cite
How Many, but How Sure? Class Competition as a Confidence Proxy in Multi-Class Density-Based Counting — arXiv:2610.23611 — arXiv (en), https://arxiv.org/abs/2610.23611 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e62f84e1e8ff (Metraipolis archive dc_e62f84e1e8ff, sha256 4e253c3663085f3239aada64dab21cdb81c03115d54a78d9ed57d9025d051cbd)
BibTeX
@misc{dce62f84e1e8ff,
title = {How Many, but How Sure? Class Competition as a Confidence Proxy in Multi-Class Density-Based Counting — arXiv:2610.23611},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.23611},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e62f84e1e8ff; sha256 4e253c3663085f3239aada64dab21cdb81c03115d54a78d9ed57d9025d051cbd},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.23611. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Multi-class density estimators can count objects in crowded scenes, but their class-wise outputs offer no indication of how reliably density mass has been allocated. We introduce a training-free confidence proxy based on local competition between the strongest class channels. The resulting prominence measure requires only the predicted density maps, with no retraining or access to model internals. We test whether residual competition is diagnostic of cross-class leakage by attenuating contested density at inference time. On DSACA predictions for the Hicks et al. dataset, this reduces overall MAE from 0.920 to 0.846, suggesting that class competition contains useful information about count reliability. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.23611 — 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 2026-10-19T09:06:29.558Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).