AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Learning Separable Fine-Grained Representation via Dendrogram Construction from Coarse Labels for Fine-grained Visual Recognition

Guanghui Shi, Xuefeng Liang, Wenjie Li, Xiaoyu Lin

ICCV 2025 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use dendrogram-based BuCSFR instead of MaskConcontrastive baseline · CIFAR100R@1: 71.24 -> 73.83Improved
Use BuCSFR instead of LCRLCR · CIFAR100-hardR@1: 63.92 -> 70.04Improved
Use BuCSFR instead of LCRLCR · FGVC-AircraftR@1: 51.79 -> 57.97Improved
Use BuCSFR instead of LCRLCR · ImageNet-1KkNN@5: 31.01 -> 33.37Improved
Use BuCSFR instead of MaskCon, training at species levelMaskCon · iNaturalist-2019 (species)R@5: 69.33 -> 76.84Improved
Use BuCSFR instead of MaskCon at genus levelMaskCon · iNaturalist-2019 (genus)R@5: 91.06 -> 92.81Improved
Use BuCSFR instead of CE Fine on CIFAR10toy R@10 (gap with fine supervision)supervised CE Fine · CIFAR10toyR@1: 93.21 -> 97.31Hurt
Set tau_s to 0.1 rather than 0.05 (ablation)BuCSFR · CIFAR100R@5: 88.49 -> 88.93Improved
Set tau_s to 0.1 rather than 0.5 (ablation)BuCSFR · CIFAR100kNN@5: 77.03 -> 78.84Improved
Set tau to 0.2 rather than 0.05 (ablation)BuCSFR · CIFAR100R@5: 87.18 -> 88.93Improved
Set tau to 0.2 rather than 0.5 (ablation)BuCSFR · CIFAR100kNN@5: 75.42 -> 78.84Improved
Vary merging threshold T from 1.5 to 10.0 (ablation)BuCSFR · CIFAR100R@5: 88.68, 88.93, 88.68, 88.53, 88.51, 88.61No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy