Metraipolis › archive › dc_bb0d715e4b50
What Drives Hierarchy-Aware Image Retrieval? Taxonomy Alignment, Objective Choice, and Geometry — arXiv:2609.25638
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.25638
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T16:39:12.895Z
- SHA-256 of the archived text
3c5fefcd5a1b73a6034d96ab2b77ac839a93d3a69544e103cea361d138e814e7- Stable URL of this page
https://metraipolis.com/documento/dc_bb0d715e4b50
How to cite
What Drives Hierarchy-Aware Image Retrieval? Taxonomy Alignment, Objective Choice, and Geometry — arXiv:2609.25638 — arXiv (en), https://arxiv.org/abs/2609.25638 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bb0d715e4b50 (Metraipolis archive dc_bb0d715e4b50, sha256 3c5fefcd5a1b73a6034d96ab2b77ac839a93d3a69544e103cea361d138e814e7)
BibTeX
@misc{dcbb0d715e4b50,
title = {What Drives Hierarchy-Aware Image Retrieval? Taxonomy Alignment, Objective Choice, and Geometry — arXiv:2609.25638},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.25638},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bb0d715e4b50; sha256 3c5fefcd5a1b73a6034d96ab2b77ac839a93d3a69544e103cea361d138e814e7},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25638. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Foundation vision models provide strong generic representations, yet high class-level retrieval accuracy does not necessarily imply that an embedding respects a target semantic taxonomy. We study strict explicit-taxonomy image retrieval on frozen DINOv2 features and ask: when hierarchical retrieval improves, how much of the change is associated with the organization of taxonomy-aware supervision, and how much with the Euclidean-hyperbolic geometry choice? We evaluate higher levels with strict cross-class criteria that exclude finer-grained matches, and compare Euclidean and hyperbolic projections trained with taxonomy-distance regression or a taxonomy-aware supervised contrastive objective. A compute-matched 2 x 2 Geometry x Loss factorial uses the same 768-256-32 projector capacity, optimization schedule, batch order, and fixed 100-epoch budget; the Loss axis denotes the Regression-to-Taxonomy-SupCon objective-family contrast. On CUB, the objective-family contrasts in mean hierarchy mAP (strict middle/high average, excluding Class/Leaf) are +0.0487 in Euclidean space and +0.0414 in hyperbolic space, compared with geometry contrasts of +0.0102 and +0.0030. On NABirds Parent-disjoint retrieval, the corresponding objective-family contrasts are +0.0467 and +0.0440, whereas geometry contrasts are +0.0017 and -0.0009. A semantic-alignment control shows that the true taxonomy substantially outperforms a structure-preserving shuffled hierarchy, while a NABirds curvature/radius control does not support stronger negative curvature as the explanation for the observed hierarchy gains. Across the two taxonomies, the Regression-to-Taxonomy-SupCon contrasts are larger in aggregate than the evaluated geometry contrasts; semantic alignment also matters separately, while geometry remains hierarchy-dependent. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-23 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.25638 — 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 2026-10-08T15:56:18.116Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).